CN111612818A - Novel binocular vision multi-target tracking method and system - Google Patents
Novel binocular vision multi-target tracking method and system Download PDFInfo
- Publication number
- CN111612818A CN111612818A CN202010384346.1A CN202010384346A CN111612818A CN 111612818 A CN111612818 A CN 111612818A CN 202010384346 A CN202010384346 A CN 202010384346A CN 111612818 A CN111612818 A CN 111612818A
- Authority
- CN
- China
- Prior art keywords
- probability
- vehicle motion
- target
- tracking
- binocular vision
- 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.)
- Withdrawn
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000005259 measurement Methods 0.000 claims description 23
- 238000001514 detection method Methods 0.000 claims description 7
- 239000011159 matrix material Substances 0.000 claims description 7
- 238000006073 displacement reaction Methods 0.000 claims description 3
- 238000012886 linear function Methods 0.000 claims description 3
- 230000000007 visual effect Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 5
- 239000013598 vector Substances 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 3
- 239000000047 product Substances 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 239000002131 composite material Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
- G06T7/251—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments involving models
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30248—Vehicle exterior or interior
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to the technical field of multi-target tracking of automatic driving, and particularly relates to a novel binocular vision multi-target tracking method and system, wherein the novel binocular vision multi-target tracking method comprises the following steps: acquiring an image through binocular vision; acquiring a moving target according to the image; constructing a vehicle motion space and a vehicle motion model; and tracking the moving target according to the vehicle motion space and the vehicle motion model through the improved joint probability data association, realizing the multi-target tracking of the intelligent vehicle, greatly improving the automation and intelligence level of a driving system, improving the tracking precision and speed, and generating no obvious deviation and missing the tracking of pedestrians when tracking the vehicle.
Description
Technical Field
The invention belongs to the technical field of multi-target tracking of automatic driving, and particularly relates to a novel binocular vision multi-target tracking method and system.
Background
Achieving reliable perception of the surrounding environment under a variety of uncertain conditions is a fundamental task in almost any assistive or autonomous system application, and especially with the continued rise of automated driving research, academia and various major technology companies are actively developing advanced driving assistance systems. The core technology of the driving assistance system comprises functions of self-adaptive cruise, collision avoidance, lane change assistance, traffic sign identification, parking assistance and the like, and aims to realize full automation of vehicle driving and reduce human errors causing road accidents while improving safety. In various technologies, moving object tracking is a key task of a driving assistance system, and when a vehicle can detect a dynamic object in its environment and predict its future behavior, it can greatly improve the level of intelligence of the vehicle.
Because real-time and accurate tracking of various objects under different environmental conditions needs to be realized, no sensing system can completely provide all information required by target tracking at present. In view of this, the driving assistance system generally achieves accurate detection of a moving target by means of a composite sensing system including a millimeter wave radar, a laser range finder, a vision system, and the like. The radar apparatus can accurately measure the relative speed and distance of an object. Laser rangefinders have a higher lateral resolution than radar and, in addition to accurately detecting object distances, can detect the footprint of objects and provide a detailed representation of the scene. The vision-based sensing system can provide accurate lateral measurement and rich image information, thereby providing an effective supplement for distance-measuring-based sensor road scene analysis. Among other things, stereo vision sensors can provide object detection with high lateral resolution and a small range of certainty, while generally providing sufficient information for the identification and classification of objects.
No matter which sensor is used, the problem of multi-target tracking must be solved in traffic scenarios. At the moment, the state of each target needs to be tracked, and meanwhile, the measured value is processed in a cluttered environment, so that the problem of data association of the tracked target is solved.
Therefore, based on the above technical problems, a new binocular vision multi-target tracking method and system need to be designed.
Disclosure of Invention
The invention aims to provide a novel binocular vision multi-target tracking method and system.
In order to solve the technical problem, the invention provides a novel binocular vision multi-target tracking method, which comprises the following steps:
acquiring an image through binocular vision;
acquiring a moving target according to the image;
constructing a vehicle motion space and a vehicle motion model; and
and tracking the moving target according to the vehicle motion space and the vehicle motion model through the improved joint probability data association.
Further, the method for acquiring the moving target according to the image comprises the following steps:
and after the image is corrected, estimating the displacement of the intermediate vehicle according to a visual stereo distance measurement algorithm to obtain a moving target.
Further, the method for constructing the vehicle motion space comprises the following steps:
constructing a vehicle motion space GL into a Cartesian product S of two matrix Lie groups S based on an equivalent principle of a rigid body constant velocity motion model1×S2；
Wherein: s1Is a position component; s2Is the velocity component.
Further, the method for constructing the vehicle motion model comprises the following steps:
building and updating vehicle motion models, i.e.
The vehicle motion model is as follows:
Xk+1＝Xk·exp(αk+βk)；
wherein, Xk∈ GL, the motion state of the system at the time k, αkAs a non-linear function βkIs white gaussian noise;
when the posterior distribution of the step k-1 meets the Gaussian distribution on the Lie group, the method is based on
Xk+1＝Xkexp(log(αk) Predicting a motion state indicated by the vehicle motion model to update the vehicle motion model to:
wherein v is1k、v2k、ωkRespectively longitudinal, transverse and rotational speed β1k、β2k、βωkThe components of the Gaussian noise in the longitudinal direction, the transverse direction and the rotation direction are respectively; t is the transposed sign of the matrix.
Further, the method for tracking the moving target according to the vehicle motion space and the vehicle motion model through the improved joint probability data association comprises the following steps:
when the number of moving objects is k, then the plurality of moving objects are represented as { T }1,...,Tn}；
YkSet representing all detections at time k
Y1:kHistory Y representing all metrics1:k＝{Y1,…,Yk}；
Predicting each moving target T according to joint probability data associationiPosterior density of (i ═ 1, 2.., n), i.e.
Wherein,
density of the target state;
density of presence of a target; p is the probability;
density of target state
And their existence
Is all YkMeasuring (2);
describing the probability of the presence of a moving object according to a Markov chain model, i.e.
Wherein: rho is the probability that the moving target still exists at the moment k when the moving target exists at the moment k-1;
from the measured value YkPredicting k time scanning to moving target T through total probability formulaiThe posterior density of (a):
wherein,
associating probabilities with posterior data of the existence of the object;
is a probability hypothesis;
the probability of the existence of the detected target is as follows:
when in use
When composed of all joint events F, with each trajectory having zero or one measurement, and each measurement is assigned to zero or one trajectory, then
TiExist ofBut not detected by the measurements within the cluster are:
wherein,
probability hypothesis for 0 measurements;
probability of 0 measurements;
p (F | Y) corresponding to each joint event F is calculated1:k) Then, a target set of measurements C is assigned, and a set of measured orbit sets D is assigned, to obtain:
wherein i is the ith moving target; pdIs TiA probability of being detected; pqFor the modified metric at TiA probability within a threshold range of (d);
is the rotation probability; rhokDensity is measured for clutter priors; τ is assigned to T at Joint event FiAn index of the measure of (d);
the rotation probability is:
the probability of the target being present is:
then the joint data probability is:
and realizing accurate tracking of a plurality of moving targets according to the joint data probability.
On the other hand, the invention also provides a novel binocular vision multi-target tracking system, which comprises:
the acquisition module acquires images through binocular vision;
the moving target acquisition module acquires a moving target according to the image;
the building module is used for building a vehicle motion space and a vehicle motion model; and
and the tracking module tracks the moving target according to the vehicle motion space and the vehicle motion model through the improved joint probability data association.
The invention has the beneficial effects that the invention obtains images through binocular vision; acquiring a moving target according to the image; constructing a vehicle motion space and a vehicle motion model; and tracking the moving target according to the vehicle motion space and the vehicle motion model through the improved joint probability data association, realizing the multi-target tracking of the intelligent vehicle, greatly improving the automation and intelligence level of a driving system, improving the tracking precision and speed, and generating no obvious deviation and missing the tracking of pedestrians when tracking the vehicle.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flow chart of a novel binocular vision multi-target tracking method according to the present invention;
FIG. 2 is a detailed flow chart of the novel binocular vision multi-target tracking method according to the present invention;
fig. 3 is a schematic block diagram of the novel binocular vision multi-target tracking system according to the present invention.
Detailed Description
To make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
Fig. 1 is a flow chart of a novel binocular vision multi-target tracking method according to the present invention.
As shown in fig. 1, this embodiment 1 provides a novel binocular vision multi-target tracking method, including: acquiring an image through binocular vision; acquiring a moving target according to the image; constructing a vehicle motion space and a vehicle motion model; and tracking the moving target according to the vehicle motion space and the vehicle motion model through improved Joint Probabilistic Data Association (JPDA), realizing intelligent vehicle multi-target tracking, greatly improving the automation and intelligence level of a driving system, improving the tracking precision and speed, and generating no obvious offset and missing the tracking of pedestrians when tracking the vehicle.
Fig. 2 is a specific flowchart of the novel binocular vision multi-target tracking method according to the present invention.
As shown in fig. 2, in the present embodiment, the method for acquiring an image through binocular vision includes: the stereoscopic vision camera is used for collecting images and videos of vehicles and pedestrians; modeling the uncertainty of the sensor under the Lie group, and performing state filtering on the preprocessed image by adopting a Euclidean group algorithm; false detections are removed using binocular vision in areas where vehicles may be present to correct the images.
In this embodiment, the method for acquiring a moving object according to an image includes: after the image is corrected, the measured uncertainty and the track of the predicted target motion are confirmed through a Kalman filter, the displacement of an intermediate vehicle is estimated by using a visual stereo distance measurement algorithm, an object which does not conform to the characteristics is regarded as a moving target, and the moving target needs to be subjected to stereo detection in the process, wherein the specific process is as follows:
firstly, after correcting the image, projecting all the feature points from the previous frame into a 3D world frame through a standard pinhole camera model, then carrying out composite back projection on the position and the obtained motion matrix into the current camera frame, and connecting to the 3D points from the current frame corresponding to the previous frame to form a vector field, wherein each vector represents the motion of the corresponding 3D point relative to the world frame; secondly, since the uncertainty of the measurement in the 3D space has a high anisotropy, it is difficult to accurately determine the intensity of motion along the optical axis direction, and the vectors are projected onto an image plane in which the uncertainty is uniformly distributed, and a threshold value is assigned to the motion amplitude of each point, and then the remaining vectors are connected into clusters according to the translation and rotation parameters; finally, if at least 3 vectors are present therein, each cluster corresponds to a moving object (moving target) and the moving object is described by the centroid point of all the corresponding points.
In this embodiment, a state uncertainty representation and a motion model are constructed based on the extended kalman filter in the Lie group, that is, a vehicle motion space and a vehicle motion model need to be constructed.
In this embodiment, the method for constructing the vehicle motion space includes: the vehicle is a typical rigid body, so the state of the vehicle needs to be described by using a rigid body motion equation set; furthermore, it is also possible to use when considering the speed of the vehicleExpressing the state change of high order by the same motion equation system; constructing a vehicle motion space GL into a Cartesian product S of two matrix Lie groups S based on an equivalent principle of a rigid body constant velocity motion model1×S2(ii) a Wherein: s1Is a position component; s2Is the velocity component.
In this embodiment, the method for constructing the vehicle motion model includes:
building and updating vehicle motion models, i.e.
The vehicle motion model is as follows:
Xk+1＝Xk·exp(αk+βk)；
wherein, Xk∈ GL, the motion state of the system at the time k, αkAs a non-linear function βkIs white gaussian noise;
if the posterior distribution of the k-1 step satisfies the Gaussian distribution on the Lie group, then X can be usedk+1＝Xkexp(log(αk) Predicting a motion state indicated by the vehicle motion model to update the vehicle motion model (re-modeling an equation of the vehicle motion model) as:
wherein v is1k、v2k、ωkRespectively longitudinal, transverse and rotational speed β1k、β2k、βωkThe components of the Gaussian noise in the longitudinal direction, the transverse direction and the rotation direction are respectively; t is the transposed sign of the matrix.
In this embodiment, the method for tracking a moving object according to a vehicle motion space and a vehicle motion model by improved joint probability data association includes:
assuming that the number of moving objects is k, the moving objects are expressed as { T }1,...,TnThe number k of moving objects to be tracked varies with time, that is to say moving objects may appear or disappear from the field of view of the sensor at any time;
definition of YkRepresenting the set of all detections at time k, i.e.
Definition of Y1:kRepresenting the history of all metrics, i.e.
Y1:k＝{Y1,…,Yk} (2)；
Predicting (estimating) each moving target T based on joint probability data associationiThe posterior density of (i ═ 1, 2.., n) solves this problem, i.e.
Wherein,
density (probability) of the target state;
density (likelihood) of presence of a target; p is the probability;
the density of the target state is expressed by the formula (3)
And their existence
(density of existence of target) is all YkAnd is related to k;
describing the probability of the presence of a moving object according to a Markov chain model, i.e.
Wherein: rho is the probability that the moving target still exists at the moment k when the moving target exists at the moment k-1;
from the measured value YkPredicting k time scanning to moving target T through total probability formulaiThe posterior density of (a) is:
wherein,
associating probabilities with posterior data of the existence of the object;
is a probability hypothesis; n iskIs a number of
The probability of the existence of the detected target is as follows:
to calculate
The associated events that measure the objects need to be considered in the set of objects; it is assumed at this time
Consisting of all joint events F, where each trajectory has zero or one measurement, and each measurement is assigned to zero or one trajectory, then
TiThe probability of being present but not detected by the measurements within the cluster is:
wherein,
probability hypothesis for 0 measurements;
probability of 0 measurements; for calculating P (F | Y) corresponding to each joint event F1:k) Then, a target set of measurements C needs to be assigned, and a set of measured trajectory sets D needs to be assigned, and then:
wherein i is the ith moving target; pdIs TiA probability of being detected; pqFor the modified metric at TiA probability within a threshold range of (d);
is the rotation probability; rhokDensity is measured for clutter priors; τ is assigned to T at Joint event FiAn index of the measure of (d);
the rotation probability can be calculated from equation (10) as:
the probability that all elements can be obtained to determine the existence of the target is given by the following public expression:
then the joint data probability is:
and realizing accurate tracking of a plurality of moving targets according to the joint data probability.
Example 2
Fig. 3 is a schematic block diagram of the novel binocular vision multi-target tracking system according to the present invention.
As shown in fig. 3, on the basis of embodiment 1, this embodiment 2 further provides a novel binocular vision multi-target tracking system, including: the acquisition module acquires images through binocular vision; the moving target acquisition module acquires a moving target according to the image; the building module is used for building a vehicle motion space and a vehicle motion model; and the tracking module tracks the moving target according to the vehicle motion space and the vehicle motion model through the improved joint probability data association.
In the embodiment, the acquisition module acquires an image through binocular vision, the moving target acquisition module acquires a moving target according to the image, and the construction module constructs a vehicle motion space and a vehicle motion model; and the method for tracking the moving target by the tracking module according to the vehicle motion space and the vehicle motion model through the improved joint probability data association has already been explained in detail in embodiment 1, and is not described in detail in this embodiment.
In summary, the invention acquires images through binocular vision; acquiring a moving target according to the image; constructing a vehicle motion space and a vehicle motion model; and tracking the moving target according to the vehicle motion space and the vehicle motion model through the improved joint probability data association, realizing the multi-target tracking of the intelligent vehicle, greatly improving the automation and intelligence level of a driving system, improving the tracking precision and speed, and generating no obvious deviation and missing the tracking of pedestrians when tracking the vehicle.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative, and for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
In addition, the functional modules in the embodiments of the present invention 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.
The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In light of the foregoing description of the preferred embodiment of the present invention, many modifications and variations will be apparent to those skilled in the art without departing from the spirit and scope of the invention. The technical scope of the present invention is not limited to the content of the specification, and must be determined according to the scope of the claims.
Claims (6)
1. A novel binocular vision multi-target tracking method is characterized by comprising the following steps:
acquiring an image through binocular vision;
acquiring a moving target according to the image;
constructing a vehicle motion space and a vehicle motion model; and
and tracking the moving target according to the vehicle motion space and the vehicle motion model through the improved joint probability data association.
2. The novel binocular vision multi-target tracking method of claim 1,
the method for acquiring the moving target according to the image comprises the following steps:
and after the image is corrected, estimating the displacement of the intermediate vehicle according to a visual stereo distance measurement algorithm to obtain a moving target.
3. The novel binocular vision multi-target tracking method of claim 2,
the method for constructing the vehicle motion space comprises the following steps:
based on the equivalent principle of the rigid body constant velocity motion model, the vehicle motion space GL is constructed into two matrix Lie group Cartesian products S1×S2；
Wherein: s1Is a position component; s2Is the velocity component.
4. The novel binocular vision multi-target tracking method of claim 3,
the method for constructing the vehicle motion model comprises the following steps:
building and updating vehicle motion models, i.e.
The vehicle motion model is as follows:
Xk+1＝Xk·exp(αk+βk)；
wherein, Xk∈ GL, which is the motion state at the time k, αkAs a non-linear function βkIs white gaussian noise;
when the posterior distribution of the step k-1 meets the Gaussian distribution on the Lie group, according to Xk+1＝Xkexp(log(αk) Predicting a motion state indicated by the vehicle motion model to update the vehicle motion model to:
wherein v is1k、v2k、ωkRespectively longitudinal, transverse and rotational speed β1k、β2k、βωkThe components of the Gaussian noise in the longitudinal direction, the transverse direction and the rotation direction are respectively; t is the transposed sign of the matrix.
5. The novel binocular vision multi-target tracking method of claim 4,
the method for tracking the moving target according to the vehicle motion space and the vehicle motion model through the improved joint probability data association comprises the following steps:
when the number of moving objects is n, then the plurality of moving objects are represented as { T }1,...,Tn}；
YkSet representing all detections at time k
Y1:kHistory Y representing all metrics1:k＝{Y1,…,Yk}；
Predicting each moving target T according to joint probability data associationiPosterior density of (i ═ 1, 2.., n), i.e.
Wherein,
density of the target state;
density of presence of a target; p is the probability;
density of target state
And their existence
Is all YkMeasuring (2);
describing the probability of the presence of a moving object according to a Markov chain model, i.e.
Wherein: rho is the probability that the moving target still exists at the moment k when the moving target exists at the moment k-1;
from the measured value YkPredicting k time scanning to moving target T through total probability formulaiThe posterior density of (a):
wherein,
associating probabilities with posterior data of the existence of the object;
is a probability hypothesis;
the probability of the existence of the detected target is as follows:
when in use
When composed of all joint events F, with each trajectory having zero or one measurement, and each measurement is assigned to zero or one trajectory, then
TiThe probability of being present but not detected by the measurements within the cluster is:
wherein,
probability hypothesis for 0 measurements;
probability of 0 measurements;
p (F | Y) corresponding to each joint event F is calculated1:k) Then, a target set of measurements C is assigned, and a set of measured orbit sets D is assigned, to obtain:
wherein i is the ith moving target; pdIs TiA probability of being detected; pqFor the modified metric at TiA probability within a threshold range of (d);
is the rotation probability; rhokDensity is measured for clutter priors; τ is assigned to T at Joint event FiAn index of the measure of (d);
the rotation probability is:
the probability of the target being present is:
then the joint data probability is:
and realizing accurate tracking of a plurality of moving targets according to the joint data probability.
6. A novel binocular vision multi-target tracking system is characterized by comprising:
the acquisition module acquires images through binocular vision;
the moving target acquisition module acquires a moving target according to the image;
the building module is used for building a vehicle motion space and a vehicle motion model; and
and the tracking module tracks the moving target according to the vehicle motion space and the vehicle motion model through the improved joint probability data association.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010384346.1A CN111612818A (en) | 2020-05-07 | 2020-05-07 | Novel binocular vision multi-target tracking method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010384346.1A CN111612818A (en) | 2020-05-07 | 2020-05-07 | Novel binocular vision multi-target tracking method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111612818A true CN111612818A (en) | 2020-09-01 |
Family
ID=72196782
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010384346.1A Withdrawn CN111612818A (en) | 2020-05-07 | 2020-05-07 | Novel binocular vision multi-target tracking method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111612818A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112304250A (en) * | 2020-10-15 | 2021-02-02 | 天目爱视（北京）科技有限公司 | Three-dimensional matching equipment and method between moving objects |
| US20220135074A1 (en) * | 2020-11-02 | 2022-05-05 | Waymo Llc | Classification of objects based on motion patterns for autonomous vehicle applications |
| CN118115755A (en) * | 2024-04-28 | 2024-05-31 | 四川盎智未来科技有限公司 | Multi-target tracking method, system and storage medium |
| US12050267B2 (en) | 2020-11-09 | 2024-07-30 | Waymo Llc | Doppler-assisted object mapping for autonomous vehicle applications |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102799900A (en) * | 2012-07-04 | 2012-11-28 | 西南交通大学 | Target tracking method based on supporting online clustering in detection |
| CN109447121A (en) * | 2018-09-27 | 2019-03-08 | 清华大学 | A kind of Visual Sensor Networks multi-object tracking method, apparatus and system |
-
2020
- 2020-05-07 CN CN202010384346.1A patent/CN111612818A/en not_active Withdrawn
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102799900A (en) * | 2012-07-04 | 2012-11-28 | 西南交通大学 | Target tracking method based on supporting online clustering in detection |
| CN109447121A (en) * | 2018-09-27 | 2019-03-08 | 清华大学 | A kind of Visual Sensor Networks multi-object tracking method, apparatus and system |
Non-Patent Citations (1)
| Title |
|---|
| 张琦等: "Lie群下利用改进JPDA 滤波器的智能车立体视觉多目标跟踪方法" * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112304250A (en) * | 2020-10-15 | 2021-02-02 | 天目爱视（北京）科技有限公司 | Three-dimensional matching equipment and method between moving objects |
| US20220135074A1 (en) * | 2020-11-02 | 2022-05-05 | Waymo Llc | Classification of objects based on motion patterns for autonomous vehicle applications |
| US12050267B2 (en) | 2020-11-09 | 2024-07-30 | Waymo Llc | Doppler-assisted object mapping for autonomous vehicle applications |
| CN118115755A (en) * | 2024-04-28 | 2024-05-31 | 四川盎智未来科技有限公司 | Multi-target tracking method, system and storage medium |
| CN118115755B (en) * | 2024-04-28 | 2024-06-28 | 四川盎智未来科技有限公司 | Multi-target tracking method, system and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3745158B1 (en) | Methods and systems for computer-based determining of presence of dynamic objects | |
| CN105892471B (en) | Automatic driving method and apparatus | |
| CN112700470B (en) | Target detection and track extraction method based on traffic video stream | |
| Wang et al. | Robust road modeling and tracking using condensation | |
| CN111612818A (en) | Novel binocular vision multi-target tracking method and system | |
| CN112132896B (en) | Method and system for detecting states of trackside equipment | |
| Erbs et al. | Moving vehicle detection by optimal segmentation of the dynamic stixel world | |
| CN110738121A (en) | front vehicle detection method and detection system | |
| CN110794406A (en) | Multi-source sensor data fusion system and method | |
| CN114758504B (en) | Online vehicle overspeed early warning method and system based on filtering correction | |
| CN114495064A (en) | Monocular depth estimation-based vehicle surrounding obstacle early warning method | |
| CN116310679A (en) | Multi-sensor fusion target detection method, system, medium, equipment and terminal | |
| CN114296095A (en) | Method, device, vehicle and medium for extracting effective target of automatic driving vehicle | |
| Zhang et al. | A novel vehicle reversing speed control based on obstacle detection and sparse representation | |
| Muresan et al. | Multimodal sparse LIDAR object tracking in clutter | |
| Hartmann et al. | Towards autonomous self-assessment of digital maps | |
| An et al. | Multi-object tracking based on a novel feature image with multi-modal information | |
| US20220309776A1 (en) | Method and system for determining ground level using an artificial neural network | |
| US20220284623A1 (en) | Framework For 3D Object Detection And Depth Prediction From 2D Images | |
| CN113511194A (en) | Longitudinal collision avoidance early warning method and related device | |
| Christiansen et al. | Monocular vehicle distance sensor using HOG and Kalman tracking | |
| Schilling et al. | Mind the gap-a benchmark for dense depth prediction beyond lidar | |
| CN115471526A (en) | Automatic driving target detection and tracking method based on multi-source heterogeneous information fusion | |
| Zeisler et al. | Analysis of the performance of a laser scanner for predictive automotive applications | |
| CN118259312B (en) | Laser radar-based vehicle collision early warning method |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20200901 |
|
| WW01 | Invention patent application withdrawn after publication |