CN115827925A - Target association method and device, electronic equipment and storage medium - Google Patents
Target association method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN115827925A CN115827925A CN202310139496.XA CN202310139496A CN115827925A CN 115827925 A CN115827925 A CN 115827925A CN 202310139496 A CN202310139496 A CN 202310139496A CN 115827925 A CN115827925 A CN 115827925A
- Authority
- CN
- China
- Prior art keywords
- target
- association
- node
- weight
- data
- 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
- 238000000034 method Methods 0.000 title claims abstract description 55
- 238000001514 detection method Methods 0.000 claims abstract description 31
- 238000012545 processing Methods 0.000 claims abstract description 28
- 238000004590 computer program Methods 0.000 claims description 15
- 230000000877 morphologic effect Effects 0.000 claims description 11
- 239000000126 substance Substances 0.000 claims description 3
- 238000010586 diagram Methods 0.000 abstract description 14
- 230000000694 effects Effects 0.000 abstract description 4
- 230000000875 corresponding effect Effects 0.000 description 31
- 238000004422 calculation algorithm Methods 0.000 description 11
- 239000003550 marker Substances 0.000 description 11
- 238000004891 communication Methods 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 7
- 230000000007 visual effect Effects 0.000 description 6
- 230000009467 reduction Effects 0.000 description 5
- 230000003190 augmentative effect Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 239000013598 vector Substances 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 102100034112 Alkyldihydroxyacetonephosphate synthase, peroxisomal Human genes 0.000 description 1
- 101000799143 Homo sapiens Alkyldihydroxyacetonephosphate synthase, peroxisomal Proteins 0.000 description 1
- 238000000848 angular dependent Auger electron spectroscopy Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a target association method, a target association device, electronic equipment and a storage medium. The method comprises the steps of obtaining first data collected by a first type sensor in a vehicle, obtaining a first target based on first data identification, obtaining second data collected by a second type sensor in the vehicle, obtaining a second target based on second data identification, generating a target association diagram based on the target data of the first target and the target data of the second target, wherein the target association diagram comprises target nodes corresponding to the first target and the second target respectively and edges between the target nodes, the target nodes and the edges are provided with weights respectively, and performing association processing on the first target and the second target based on the weights corresponding to the target nodes and the edges in the target association diagram respectively to obtain a detection target of the vehicle. The method and the device can ensure a better and more stable correlation effect of the target under the condition of ensuring the detection characteristic advantages of each sensor.
Description
Technical Field
The embodiment of the invention relates to the field of automatic driving perception fusion, in particular to a target association method, a target association device, electronic equipment and a storage medium.
Background
At present, the automatic driving technology is developed rapidly, and the medium-high end vehicle types at home and abroad have the iterative intelligent driving function. On a mass production vehicle with an ADAS function, common sensors such as a laser radar, a camera and a millimeter wave radar are usually equipped as target sensing sources, and for the result of multi-source heterogeneous acquisition, target-level fusion is usually adopted, and the same target information from different sensing sources is subjected to correlation matching and post-processing.
Common data association methods include a nearest neighbor matching algorithm (NN), a joint probability data association algorithm (JPDA), a multiple hypothesis tracking algorithm (MHT), and a hungarian matching algorithm (Hungary). The NN algorithm is high in instantaneity, but is frequently subjected to false association when encountering a complex environment, so that risks are caused; the JPDA and MHT are high in correlation accuracy in a target-intensive environment, but the calculated amount is large, the real-time performance is difficult to meet, and the traditional Hungary algorithm is maximum matching and cannot guarantee optimal matching.
Disclosure of Invention
The invention provides a target association method, a target association device, electronic equipment and a storage medium, which are used for realizing a better target association effect.
In a first aspect, an embodiment of the present invention provides a target association method, including:
acquiring first data acquired by a first sensor in a vehicle, identifying to obtain a first target based on the first data, acquiring second data acquired by a second sensor in the vehicle, and identifying to obtain a second target based on the second data;
generating a target association graph based on target data of the first target and target data of the second target, wherein the target association graph comprises target nodes corresponding to the first target and the second target respectively and edges between the target nodes, and the target nodes and the edges are provided with weights respectively;
and performing association processing on the first target and the second target based on the weight corresponding to each target node and each edge in the target association graph to obtain the detection target of the vehicle.
Optionally, generating a target association map based on the target data of the first target and the target data of the second target includes:
determining an association similarity of the first target and the second target based on the target data of the first target and the target data of the second target;
respectively setting a first target node corresponding to the first target and a second target node corresponding to the second target;
and setting the weight of each first target node, the weight of each second target node and the weight of an edge between each first target node and each second target node based on the association similarity of the first target and the second target to form a target association graph.
Optionally, the setting the weight of each first target node, the weight of each second target node, and the weight of the edge between each first target node and each second target node based on the associated similarity between the first target and the second target includes:
setting the weight of the edge between the first target node and the second target node based on the associated similarity of the first target and the second target;
and determining the weight of the first target node based on the association similarity between the first target and each second target, and setting the weight of each second target node as a preset value.
Optionally, the performing, based on the weights respectively corresponding to each target node and each edge in the target association graph, association processing on the first target and the second target includes:
matching the weight of a first target node with the weight of an edge connected with the first target node, determining a second target node connected with the successfully matched edge, and associating a first target corresponding to the first target node with a second node corresponding to the second target node.
Optionally, after determining the second target node connected by the successfully matched edge, the method further includes:
if at least two first target nodes are matched with the same second target node, determining the weight difference of the at least two first target nodes relative to the second target node respectively, and determining the first target node matched with the second target node based on the weight difference.
Optionally, the determining the weight difference between the at least two first target nodes and the second target node respectively includes:
for any first target node, determining the weight of the first target node, the weight of the second target node and the weight of the edge between the first target node and the second target node as the weight of the first target node relative to the second target node;
and determining the difference value of the weights of any two first target nodes relative to the second target node as the weight difference.
Optionally, the determining, based on the weight difference, a first target node that matches the second target node includes:
if the weight difference meets a threshold condition, determining a first target node matched with the second target node in at least two first target nodes;
if the weight difference does not meet the threshold condition, adjusting one or more of the weights of the at least two first target nodes, the weights of the second target nodes and the edge weights of the at least two first target nodes and the second target nodes based on weight adjusting parameters, and re-determining the first target nodes matched with the second target nodes based on the adjusted weights.
Optionally, the weight adjustment parameter is a preset fixed value; alternatively, the first and second electrodes may be,
the weight adjustment parameter is the minimum of the weights of at least two of the first target nodes relative to the second target node.
Optionally, before generating the target association map based on the target data of the first target and the target data of the second target, the method further includes:
setting a correlation range based on the first target, and determining the first target and the second target in the correlation range;
correspondingly, the generating a target association map based on the target data of the first target and the target data of the second target includes:
and for each association range, generating a target association graph based on the target data of the first target and the target data of the second target in the association range.
Optionally, after determining the first target and the second target within the association range, the method further includes:
for each association range, determining a first target number and a second target number in the association range;
if the first target number is greater than one and/or the second target number is greater than one, continuing to execute a step of generating a target association graph based on the target number of the first target and the target number of the second target in the association range;
if the first target number is less than one, or the second target number is less than one, determining the second target or the first target as a detection target;
and if the number of the first targets is equal to one and the number of the second targets is equal to one, performing association processing on the first targets and the second targets based on the association similarity of the first targets and the second targets to determine detection targets.
Optionally, the performing, based on the association similarity between the first target and the second target, association processing on the first target and the second target to determine a detection target includes:
determining motion similarity and/or morphological similarity of the first target and the second target based on the target data of the first target and the target data of the second target;
and judging the motion similarity and/or the form similarity of the first target and the second target based on the association threshold, and determining the association relation of the first target and the second target.
In a second aspect, an embodiment of the present invention further provides a target association apparatus, including:
the target identification module is used for acquiring first data acquired by a first type of sensor in a vehicle, identifying to obtain a first target based on the first data, acquiring second data acquired by a second type of sensor in the vehicle, and identifying to obtain a second target based on the second data;
an association graph generation module, configured to generate a target association graph based on target data of the first target and target data of the second target, where the target association graph includes target nodes corresponding to the first target and the second target, and edges between the target nodes, and the target nodes and the edges are provided with weights, respectively;
and the association processing module is used for performing association processing on the first target and the second target based on the weights corresponding to the target nodes and the edges in the target association diagram respectively to obtain the detection target of the vehicle.
In a third aspect, an embodiment of the present invention further provides an electronic device, where the electronic device includes:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the target association method of any one of the first aspect.
In a fourth aspect, the present invention further provides a computer-readable storage medium, which stores computer instructions for causing a processor to implement the target association method in any one of the first aspect when executed.
According to the method, first data collected by a first type of sensor in the vehicle are obtained, a first target is obtained based on first data identification, second data collected by a second type of sensor in the vehicle are obtained, a second target is obtained based on second data identification, a target association diagram is generated based on the target data of the first target and the target data of the second target, the target association diagram comprises target nodes corresponding to the first target and the second target respectively and edges between the target nodes, the target nodes and the edges are provided with weights respectively, and the first target and the second target are subjected to association processing based on the weights corresponding to the target nodes and the edges in the target association diagram respectively, so that the detection target of the vehicle is obtained. The method and the device can ensure a better and more stable correlation effect of the target under the condition of ensuring the detection characteristic advantages of each sensor.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the description below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a flowchart of a target association method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a target association graph according to an embodiment of the invention;
fig. 3 is a flowchart of a target node matching method according to an embodiment of the present invention;
fig. 4 is a flowchart of a target association method according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a target association apparatus according to a second embodiment of the present invention;
fig. 6 is a schematic structural diagram of an electronic device according to a third embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, shall fall within the protection scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
Example one
Fig. 1 is a flowchart of a target association method according to an embodiment of the present invention, where the embodiment is applicable to a target association situation, and the method may be executed by a target association apparatus, where the target association apparatus may be implemented in a form of hardware and/or software, and the target association apparatus may be configured in an electronic device such as a computer, a server, a mobile terminal, and the like. As shown in fig. 1, the method includes:
s110, first data collected by a first type of sensor in the vehicle are obtained, a first target is obtained based on the first data in a recognition mode, second data collected by a second type of sensor in the vehicle are obtained, and a second target is obtained based on the second data in a recognition mode.
The sensor may be a detection device, and the sensor here may be a sensor arranged in a vehicle, and is used for checking environment information of the vehicle, and converting the detected information into an electric signal or other information output in a required form according to a certain rule, so as to meet requirements of information transmission, processing, storage, display, recording, control and the like, for example, a laser radar, a camera, an ultrasonic device and the like. For example, a laser radar may be used to detect point cloud data of the environment in which the vehicle is located, a camera may be used to detect image data of the environment in which the vehicle is located, and so on. The data detected by each sensor may be processed by a preset algorithm to determine a target in the environment, where the detected target may be a target to be referred to by the vehicle during driving, and the target includes, but is not limited to, other vehicles, pedestrians, obstacles, and the like in the road. It can be understood that data detected by different types of sensors can be processed based on different preset algorithms, different types of targets can be detected based on different preset algorithms, the specific type of the preset algorithm is not limited here, and any type of data detected by any type of sensors can be used as long as each type of target is identified. Wherein the data detected by each type of sensor can identify a plurality of targets, where the types of targets can be the same or different.
The first target may be identified by data collected by the first sensor and the second target may be identified by data collected by the second sensor. And determining target level state vectors of the first target and the second target according to the target identification result. The target level state vector may include, but is not limited to, target lateral-to-longitudinal distance, lateral-to-longitudinal relative velocity, target category, heading angle, size, and the like. In some embodiments, the data collected by the sensor may be transmitted to a pre-trained target recognition model, and a target-level state vector of each target output by the target recognition model is obtained.
Optionally, after first data acquired by a first type of sensor and second data acquired by a second type of sensor in the vehicle are acquired, time-space synchronization is performed on the first data and the second data. The time synchronization may be to synchronize the point cloud data with a time stamp of the road image, for example, an in-vehicle terminal system or an inertial navigation UTC provides reference time information, distributes the reference time information to each sensor in the vehicle, and accordingly, each sensor sets a time stamp for the acquired data based on the reference time information. The spatial synchronization may be to convert the point cloud data and the road image into the same coordinate system, where the coordinate system may be a preset coordinate system, and preset a coordinate conversion relationship between each sensor and the preset coordinate system, and the acquired point cloud data and the road image are converted based on the corresponding coordinate conversion relationship, respectively, to obtain the point cloud data and the road image in the same coordinate system, thereby completing the spatial synchronization.
It should be noted that the recognition accuracy of the first type of sensor may be higher than that of the second type of sensor. For example, the first type of sensor may be a lidar, the second type of sensor may be a camera, the first data may be point cloud data, and the second data may be visual image data.
In some embodiments, object detection may be performed by a variety of sensors, where objects detected by any two types of sensors are correlated. In some embodiments, for any one type of sensor, a plurality of sensors of the above type may be provided in the vehicle, and accordingly, data collected by the same type of sensor is subjected to fusion processing, and target detection is performed based on the fused data.
And S120, generating a target association graph based on the target data of the first target and the target data of the second target, wherein the target association graph comprises target nodes corresponding to the first target and the second target respectively and edges between the target nodes, and the target nodes and the edges are provided with weights respectively.
The target data of the first target is target level data of the first target, and the target data of the second target is target level data of the second target. The object association graph may be a model composed of two mutually disjoint point sets, the point sets may be object nodes, and accordingly, the first object constitutes a first object set, the second object constitutes a second object set, where the number of the first objects may be at least one, and the number of the second objects may be at least one. Further, the number of at least one of the first target and the second target is two or more. The target association graph is a bipartite graph formed by association relations between first targets and second targets, wherein each first target or second target is respectively used as a target node. The edge may be an association line associating any two target nodes in two mutually non-intersecting point sets, and is used to represent a relationship between the first target and the second target, and correspondingly, the two target nodes associated with each edge in the target association graph belong to two different point sets respectively.
In the target association graph, each target node and each edge are respectively provided with a weight, and any weight can be obtained by calculation based on target data of the first target and target data of the second target.
For example, referring specifically to FIG. 2, a first set of target points is established for two types of sensor targets, respectivelyL d And a second set of target pointsV d Form a bipartite graph and a point setL d AndV d the point in (1) is a target node, and a connecting line between each target node is an edge E of the bipartite graph. It should be noted that the number of edges and target nodes in fig. 2 is only an example, and the target nodes in the second target point set need to be traversed based on the target nodes in the first target point set during actual calculation.
Optionally, before generating the target association graph based on the target data of the first target and the target data of the second target, a matching list may be established for storing the association result, that is, the target object that needs to be focused finally. For example, for the detection results of any two sensors, if only the first object is detected and the second object is empty, the first object is added to the matching list; and if only the second target is detected and the first target is empty, adding the second target to the matching list. And in the case that at least one first target and at least one second target are detected, determining the association relationship between the first target and the second target by establishing a target association graph, and determining target data added into the matching list based on the association result.
Optionally, based on the target data of the first target and the target data of the second target, determining the association similarity between the first target and the second target, respectively setting a first target node corresponding to the first target and a second target node corresponding to the second target, and based on the association similarity between the first target and the second target, setting the weight of each first target node, the weight of each second target node, and the weight of an edge between each first target node and each second target node, to form a target association graph.
The target data may be a target-level state vector, i.e., information such as a target longitudinal-transverse distance, a longitudinal-transverse relative speed, a target category, a heading angle, and a size. The association similarity may be a result calculated in a preset manner based on the target data, and may include one or more of motion state similarity, morphological similarity, and the like. Wherein the motion state similarity may be determined based on a target lateral-to-longitudinal distance and/or a lateral-to-longitudinal relative velocity of the first target and the second target, and the morphology similarity may be determined based on a heading angle and/or a size of the first target and the second target. The correlation similarity may be calculated based on a preset distance formula, such as a mahalanobis distance formula, an euclidean distance formula, and the like, which is not limited herein.
For example, for a first target and a second target which only establish a single matching relationship, the motion state similarity of the first target and the second target is calculated according to the mahalanobis distanceD ij And calculating the morphological similarity by the Mahalanobis distance formulaT ij . Taking the motion state similarity as an example, the formula is as follows:
wherein the content of the first and second substances,D ij the motion state similarity between the jth first target node in the first target point set and the ith second target node in the second target point set,V i a distance velocity matrix of target nodes detected for the second type of sensor,L j is a range velocity matrix of the target node of the first type of sensor,P vi andP lj covariance matrices of the two types of sensors are respectively.
And calculating the association similarity of the targets based on the target data, and setting the corresponding nodes and weights of the targets to form a target association graph so as to realize association processing of the first target and the second target.
Optionally, the weight of the edge between the first target node and the second target node is set based on the association similarity between the first target and the second target, the weight of the first target node is determined based on the association similarity between the first target and each second target, and the weight of each second target node is set to be a preset value.
The identification precision of the first type of sensor can be higher than that of the second type of sensor, the weight of the second target node is set to be 0, each first target node is traversed, and the weight of the first target node is set based on the association similarity between the first target and each second target. It should be noted that, the weights of the first target node and the second target node in the target association graph are assigned based on the association similarity between the first target and the second target.
For example, referring specifically to FIG. 2, based on motion state similarityD ij Assigning a first set of target pointsL d Each first target node weightW Li =maxD ij I.e. maximum of the associated similarity between the first object and the respective second object, set of second object pointsV d Weighting with preset valueW vj =0, setting the weight of each matching edge according to the matching relationshipW Ei . Here, the value of the weight of each edge is a value of the degree of similarity in motion state of the first object and the second object. It is understood that in other embodiments, the weight of the first target node and the weight of the edge in the target association graph may be determined based on the morphological similarity; alternatively, the shape similarity and the motion similarity may be determined comprehensively, for example, based on an average value of the shape similarity and the motion similarity, and the like, which is not limited herein.
And S130, performing association processing on the first target and the second target based on the weights corresponding to each target node and each edge in the target association graph to obtain the detection target of the vehicle.
Wherein the detection target may be a target of interest to the final vehicle.
After the target association graph is generated, association processing between the first target and the second target is performed based on the target association graph, specifically, the weights of the first target nodes and the weights of the edges connected to the first target nodes are matched, the second target nodes connected to the edges successfully matched are determined, and the first target corresponding to the first target node and the second nodes corresponding to the second target nodes are associated.
For example, a first target node in the first target point set is scanned, an edge with an edge weight equal to the weight of the first target node is found, a corresponding second target node in the second target point set is determined based on the edge, the second target node and the second target node are matched, then, the second target node is matched with the next first target node in the first target point set, and the operation is repeated until all first target nodes in the first target point set are traversed.
The target nodes in the second target point set are traversed based on the target nodes in the first target point set, and the sensor with higher precision is used as the center for matching, so that the association precision is further improved. And under the condition that no conflict exists between the second target nodes matched with each first target node, determining the associated first target and second target based on the matching relation. And adding the successfully matched target data in the matching list based on the correlation processing result.
On the basis of the above embodiment, if at least two first target nodes match with the same second target node, that is, a second target node conflict that at least two first target nodes match with exists, an optimal solution is determined among the at least two first target nodes. Specifically, the weight difference between at least two first target nodes with respect to the second target node is determined, and the first target node matched with the second target node is determined based on the weight difference. It can be understood that if there are three or more first target nodes matching the same second target node, the first target nodes are compared pairwise until the optimal solution of at least two first target nodes is determined.
The weight difference may be a result calculated based on a preset manner, and is used for comparing the relationship between the target nodes. Optionally, for any first target node, the weight of the second target node, and the edge weight between the first target node and the second target node are determined as the weight of the first target node relative to the second target node, and the difference between the weights of any two first target nodes relative to the second target node is determined as the weight difference.
Optionally, if the weight difference satisfies the threshold condition, a first target node matched with the second target node is determined in the at least two first target nodes.
For example, if there are at least two first target nodesL d1 AndL d2 with the same second target nodeV d1 If the two first target nodes are matched, determining two first target nodesL d1 AndL d2 with the same second target nodeV d1 Weight difference of (3), presetTh d Is the weight difference threshold of the motion state, if comparedd D1 ＜Th d Then, thenL d1 Keeping the matching relation, refusing to update the matching relation of the target node, otherwise, comparing the two repeated matching nodesL d1 AndL d2 to pairV d1 Morphology similarity ofT ij Calculating the difference of the morphological weights
Is provided withTh T Is a form weight difference threshold, if comparedd T ＜Th T Then, thenL d1 Keeping the matching relationship, and refusing to update the matching relationship of the target node, wherein the formula is as follows:
wherein, W L1 Is the first target node L d1 Weight of (1), W V1 Is a second target node V d1 Weight of (1), W E11 Is L d1 And V d1 Weight of edges between, W L2 Is a second first target node L d2 Weight of (1), W E21 Is L d2 And V d1 The weight of the edges in between.
Optionally, if the weight difference does not satisfy the threshold condition, one or more of the weights of the at least two first target nodes, the weight of the second target node, and the edge weights of the at least two first target nodes and the second target node are adjusted based on the weight adjustment parameter, and the first target node matched with the second target node is re-determined based on the adjusted weights.
The weight adjustment parameter may be a fixed value set according to actual conditions, and is not specifically limited herein.
For example, if the weight difference does not satisfy the threshold condition, the weight adjustment parameter is setd WE Reducing the edge weight:
if the weight is poord D1 And if the threshold condition is met and at least two first target nodes are not matched with the same second target node, judging that the matching is successful and writing the matching into the association list.
Optionally, after determining the first target node matched with the second target node based on the weight difference, the new matching pointL d2 Is not provided withV d1 If the target node is optimally matched, and other target nodes capable of being matched do not exist, the target node is directly added into the association list, otherwise, the target node benchmarking value is reduced, and the formula is as follows:
wherein, ifL d1 AndL d2 are all provided withV d1 Of the optimal solution matching objectMarking nodes, calculating:
and adjusting the first target point to be concentratedL d1 ，L d1 And the second target point is concentratedV d1 The weight value of (c):
if the weight is poord D1 And if the threshold condition is met and at least two first target nodes are not matched with the same second target node, determining matching and writing the matching into a matching list.
Optionally, at least two first target nodes are matched with the same second target node, after the first target nodes matched with the second target nodes are determined based on the weight difference, all points which fail to be matched are re-matched, and if the matching times reach the upper limit and the matching is not successful, it is determined that the first target does not have the second target associated with the threshold. The matching times can be adaptively set according to actual conditions.
When at least two first target nodes are matched with the same second target node, the matching relation is updated based on the weight difference between the target nodes, so that the matching of the targets is more accurate, and the association precision is further improved.
In an alternative embodiment, and with particular reference to fig. 3, when a lidar having multiple matching relationships matches a visual target, i.e. there are at least two first target nodes matching the same second target node. And establishing a bipartite graph. And (4) assigning weights to edges and vertexes (target nodes) according to the maximum similarity of the motion states. The set of lidar points (the first set of target points) and the matching edges are traversed until a matching conflict is encountered (traversing to a second first target node to match the same second target node). And judging whether the motion state similarity difference is larger than a threshold value. If so, selecting an updated matching state with the maximum similarity, adding the updated matching state into a matching list (association list), and searching an augmented path by the route reduction marker post which fails in the matching, namely, adjusting the target node which fails in the matching based on the weight adjusting parameter and then re-matching. If not, judging whether the morphological similarity difference is larger than a threshold value. If so, selecting the updated matching state with the maximum similarity, adding the updated matching state into a matching list (association list), and searching for an augmented path by the route reduction marker post with the current matching failure. If not, reducing the laser radar track marker post (first target node weight), increasing the visual track marker post (second target node weight), and comparing whether the difference between the sum of the two track marker post values (weight values) and the edge weight value (weight) meets a threshold value. If yes, selecting the updated matching state with the minimum difference, adding the updated matching state into the matching list, and searching for an augmented path by the route reduction marker post with the current matching failure. If not, reducing the side weight, comparing whether the difference between the sum of the two track marker post values and the side weight meets a threshold value, if so, selecting the updated matching state with the minimum difference, adding the updated matching state into a matching list, and searching an augmented path for the track reduction marker post with the current matching failure. If not, judging whether the cycle number reaches the upper limit or not. If so, directly adding the laser radar track with the minimum motion state similarity into the track result, wherein the track has no matched visual track, and accumulating at other points to search for other matches. If not, returning to execute the reduction of the laser radar track marker post, increasing the visual track marker post, and comparing whether the difference between the sum of the two track marker post values and the edge weight value meets the threshold value.
Optionally, the weight adjustment parameter is a preset fixed value, or the weight adjustment parameter is a minimum value of weights of the at least two first target nodes relative to the second target node.
On the basis of any of the above embodiments, before generating the object association map based on the object data of the first object and the object data of the second object, setting an association range based on the first object, and determining the first object and the second object within the association range.
The association range may be an association box consisting of coordinate neighborhoods, which is set according to actual conditions. For example, each first target detected by the first type of sensor is used as a center, a rough position coordinate neighborhood Ω is set according to experience to form an association frame, further association matching is performed within the range of the association frame, and a matching relationship is established between the first target and a second target within the frame.
Accordingly, generating the target association graph based on the target data of the first target and the target data of the second target may include generating the target association graph based on the target data of the first target and the target data of the second target within the association range for each association range.
By setting a certain association frame area, the targets needing matching are reduced, and the calculation pressure is reduced.
Optionally, for each association range, determining a first target number and a second target number in the association range, and if the first target number is greater than one and/or the second target number is greater than one, continuing to perform the step of generating the target association map based on the target number of the first target and the target number of the second target in the association range.
The number of targets may be the number of targets recognized by each type of sensor in the association range. For example, if the first target number is 3 and the second target number is also 3, the step of generating the target association map based on the target number of the first target and the target number of the second target in the association range is continuously executed.
Optionally, if the number of the first targets is less than one, or the number of the second targets is less than one, the second targets or the first targets are determined as the detection targets. For example, if the first target number is 3 and the second target number is 0, the first target is determined as the detection target.
Optionally, if the number of the first targets is equal to one and the number of the second targets is equal to one, the first targets and the second targets are associated based on the association similarity of the first targets and the second targets, and the detection targets are determined.
For example, if the number of first targets is 1 and the number of second targets is also 1, the first targets and the second targets are associated based on the association similarity between the first targets and the second targets, and the detection targets are determined. Similarity of motion statesD ij Degree of similarity to motion stateThreshold valueTh d Performing comparison calculation, and comparing the morphological similarityT ij And morphology similarity thresholdTh m Performing comparison calculation if
And if the first target and the second target are successfully matched, correlating the first target and the second target and putting the first target and the second target into a matching list, otherwise, determining the first target and the second target as different targets and not performing matching correlation.
Subsequent execution contents are determined through the target number, excessive calculation is not needed when no matching target or a single matching target exists, and the calculation burden is reduced.
Optionally, the motion similarity and/or the morphological similarity between the first target and the second target is determined based on the target data of the first target and the target data of the second target, and the motion similarity and/or the morphological similarity between the first target and the second target are determined based on the association threshold, so as to determine the association relationship between the first target and the second target.
In an alternative embodiment, referring to fig. 4 in particular, the first data and the second data are obtained by a laser radar (first type sensor) and a camera (second type sensor) and the target data is output after the target is identified. And performing space-time synchronization on the target data. Setting a field (setting an association range based on a first target), and establishing a to-be-matched relation for the first time. And if the relation to be matched is not established, adding the result set A (association list). If the relation to be matched is successfully established, the motion state similarity and the form similarity are calculated, and whether the similarity meets the requirement or not is judged. If not, adding the result set A. If yes, judging whether the matching is single. If yes, establishing a matching relation and adding the result set A. If not, establishing a bipartite graph. And searching whether a stable optimal match exists. If yes, establishing a matching relation and adding the result set A. If not, the target is judged to be a single target, and a result set A is added.
According to the technical scheme of the embodiment, a first target is obtained through obtaining first data collected by a first type of sensor in a vehicle and based on the first data identification, second data collected by a second type of sensor in the vehicle is obtained and based on the second data identification, a target association graph is generated based on the target data of the first target and the target data of the second target, the target association graph comprises target nodes corresponding to the first target and the second target respectively and edges between the target nodes, the target nodes and the edges are provided with weights respectively, and the first target and the second target are associated and processed based on the weights corresponding to the target nodes and the edges in the target association graph, so that a detection target of the vehicle is obtained. The method and the device can ensure a better and more stable correlation effect of the target under the condition of ensuring the detection characteristic advantages of each sensor.
Example two
Fig. 5 is a schematic structural diagram of a target association apparatus according to a second embodiment of the present invention. As shown in fig. 5, the apparatus includes:
the target identification module 510 is configured to obtain first data acquired by a first type of sensor in a vehicle, identify to obtain a first target based on the first data, obtain second data acquired by a second type of sensor in the vehicle, and identify to obtain a second target based on the second data;
an association graph generating module 520, configured to generate a target association graph based on the target data of the first target and the target data of the second target, where the target association graph includes target nodes corresponding to the first target and the second target, respectively, and edges between the target nodes, and the target nodes and the edges are provided with weights, respectively;
the association processing module 530 is configured to perform association processing on the first target and the second target based on weights corresponding to each target node and each edge in the target association map, so as to obtain a detection target of the vehicle.
Optionally, the association map generating module 520 includes:
the association similarity determining module is used for determining the association similarity of the first target and the second target based on the target data of the first target and the target data of the second target;
a node setting module, configured to set a first target node corresponding to the first target and a second target node corresponding to the second target, respectively;
and the target association graph forming module is used for setting the weight of each first target node, the weight of each second target node and the weight of an edge between each first target node and each second target node based on the association similarity of the first target and the second target to form a target association graph.
Optionally, the target association map forming module is specifically configured to:
setting the weight of the edge between the first target node and the second target node based on the associated similarity of the first target and the second target;
and determining the weight of the first target node based on the association similarity between the first target and each second target, and setting the weight of each second target node as a preset value.
Optionally, the association processing module 530 includes:
and the matching and associating module is used for matching the weight of the first target node with the weight of the edge connected with the first target node, determining a second target node connected with the edge successfully matched, and associating the first target corresponding to the first target node with the second node corresponding to the second target node.
Optionally, the association processing module 530 further includes:
and the re-matching module is used for determining the weight difference of at least two first target nodes relative to the second target node respectively if at least two first target nodes are matched with the same second target node after determining the second target node connected with the successfully matched edge, and determining the first target node matched with the second target node based on the weight difference.
Optionally, the re-matching module includes:
a weight determining module, configured to determine, for any of the first target nodes, a weight of the first target node, a weight of the second target node, and an edge weight between the first target node and the second target node as a weight of the first target node relative to the second target node;
and the weight difference determining module is used for determining the difference value of the weights of any two first target nodes relative to the second target node as the weight difference.
Optionally, the re-matching module includes:
a threshold comparison module, configured to determine, if the weight difference satisfies a threshold condition, a first target node that matches the second target node among the at least two first target nodes;
and the adjusting module is used for adjusting one or more of the weights of the at least two first target nodes, the weight of the second target node and the edge weights of the at least two first target nodes and the second target node based on a weight adjusting parameter if the weight difference does not meet a threshold condition, and re-determining the first target node matched with the second target node based on the adjusted weights.
Optionally, the weight adjustment parameter is a preset fixed value; alternatively, the first and second electrodes may be,
the weight adjustment parameter is the minimum of the weights of at least two of the first target nodes relative to the second target node.
Optionally, the target associating apparatus further includes:
the target determining module is used for setting a correlation range based on the first target and determining the first target and the second target in the correlation range before generating a target correlation diagram based on target data of the first target and target data of the second target;
correspondingly, the association map generating module 520 is specifically configured to:
and for each association range, generating a target association graph based on the target data of the first target and the target data of the second target in the association range.
Optionally, the target determining module includes:
a quantity determination module for determining, for each association range, a first target quantity and a second target quantity within the association range after determining the first target and the second target within the association range;
a first quantity judgment module, configured to continue to execute the step of generating a target association graph based on the number of the first targets and the number of the second targets in the association range if the number of the first targets is greater than one and/or the number of the second targets is greater than one;
a detection target determining module, configured to determine the second target or the first target as a detection target if the number of the first targets is less than one, or the number of the second targets is less than one;
and the second quantity judging module is used for performing association processing on the first target and the second target based on the association similarity of the first target and the second target to determine a detection target if the quantity of the first target is equal to one and the quantity of the second target is equal to one.
Optionally, the second quantity determining module is specifically configured to:
determining motion similarity and/or morphological similarity of the first target and the second target based on the target data of the first target and the target data of the second target;
and judging the motion similarity and/or the form similarity of the first target and the second target based on the association threshold, and determining the association relation of the first target and the second target.
The target association device provided by the embodiment of the invention can execute the target association method provided by any embodiment of the invention, and has corresponding functional modules and beneficial effects of the execution method.
EXAMPLE III
Fig. 6 is a schematic structural diagram of an electronic device according to a third embodiment of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.
As shown in fig. 6, the electronic device 10 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a Read Only Memory (ROM) 12, a Random Access Memory (RAM) 13, and the like, wherein the memory stores a computer program executable by the at least one processor, and the processor 11 can perform various suitable actions and processes according to the computer program stored in the Read Only Memory (ROM) 12 or the computer program loaded from a storage unit 18 into the Random Access Memory (RAM) 13. In the RAM 13, various programs and data necessary for the operation of the electronic apparatus 10 can also be stored. The processor 11, the ROM12, and the RAM 13 are connected to each other via a bus 14. An input/output (I/O) interface 15 is also connected to bus 14.
A number of components in the electronic device 10 are connected to the I/O interface 15, including: an input unit 16 such as a keyboard, a mouse, or the like; an output unit 17 such as various types of displays, speakers, and the like; a storage unit 18 such as a magnetic disk, an optical disk, or the like; and a communication unit 19 such as a network card, modem, wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information/data with other devices via a computer network such as the internet and/or various telecommunication networks.
In some embodiments, the target association method may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 10 via the ROM12 and/or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the object association method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the target association method by any other suitable means (e.g., by means of firmware).
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), system on a chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.
A computer program for implementing the object association method of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be performed. A computer program can execute entirely on a machine, partly on a machine, as a stand-alone software package partly on a machine and partly on a remote machine or entirely on a remote machine or server.
Example four
A fourth embodiment of the present invention further provides a computer-readable storage medium, where computer instructions are stored, and the computer instructions are configured to enable a processor to execute a target association method, where the method includes:
acquiring first data acquired by a first type of sensor in a vehicle, identifying to obtain a first target based on the first data, acquiring second data acquired by a second type of sensor in the vehicle, and identifying to obtain a second target based on the second data;
generating a target association graph based on target data of the first target and target data of the second target, wherein the target association graph comprises target nodes corresponding to the first target and the second target respectively and edges between the target nodes, and the target nodes and the edges are provided with weights respectively;
and performing association processing on the first target and the second target based on the weight corresponding to each target node and each edge in the target association graph to obtain the detection target of the vehicle.
In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. A computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), blockchain networks, and the internet.
The computing system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical host and VPS service are overcome.
It should be understood that various forms of the flows shown above, reordering, adding or deleting steps, may be used. For example, the steps described in the present invention may be executed in parallel, sequentially, or in different orders, and are not limited herein as long as the desired result of the technical solution of the present invention can be achieved.
The above-described embodiments should not be construed as limiting the scope of the invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made, depending on design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (14)
1. An object association method, comprising:
acquiring first data acquired by a first type of sensor in a vehicle, identifying to obtain a first target based on the first data, acquiring second data acquired by a second type of sensor in the vehicle, and identifying to obtain a second target based on the second data;
generating a target association graph based on the target data of the first target and the target data of the second target, wherein the target association graph comprises target nodes corresponding to the first target and the second target respectively and edges between the target nodes, and the target nodes and the edges are provided with weights respectively;
and performing association processing on the first target and the second target based on the weight corresponding to each target node and each edge in the target association graph to obtain the detection target of the vehicle.
2. The method of claim 1, wherein generating an object association graph based on the object data for the first object and the object data for the second object comprises:
determining an association similarity of the first target and the second target based on the target data of the first target and the target data of the second target;
respectively setting a first target node corresponding to the first target and a second target node corresponding to the second target;
and setting the weight of each first target node, the weight of each second target node and the weight of an edge between each first target node and each second target node based on the association similarity of the first target and the second target to form a target association graph.
3. The method according to claim 2, wherein the setting of the weight of each first target node, the weight of the second target node, and the weight of the edge between each first target node and the second target node based on the associated similarity of the first target and the second target comprises:
setting the weight of the edge between the first target node and the second target node based on the associated similarity of the first target and the second target;
and determining the weight of the first target node based on the association similarity between the first target and each second target, and setting the weight of each second target node as a preset value.
4. The method according to claim 2, wherein the associating the first object and the second object based on the weight corresponding to each object node and each edge in the object association graph comprises:
matching the weight of a first target node with the weight of an edge connected with the first target node, determining a second target node connected with the edge successfully matched, and associating a first target corresponding to the first target node with a second node corresponding to the second target node.
5. The method of claim 4, after determining the second target node connected by the edge successfully matched, further comprising:
if at least two first target nodes are matched with the same second target node, determining the weight difference of the at least two first target nodes relative to the second target node respectively, and determining the first target node matched with the second target node based on the weight difference.
6. The method of claim 5, wherein determining the respective weight differences of the at least two first target nodes relative to the second target node comprises:
for any first target node, determining the weight of the first target node, the weight of the second target node and the weight of the edge between the first target node and the second target node as the weight of the first target node relative to the second target node;
and determining the difference value of the weights of any two first target nodes relative to the second target node as the weight difference.
7. The method of claim 6, wherein determining the first target node that matches the second target node based on the weight difference comprises:
if the weight difference meets a threshold condition, determining a first target node matched with the second target node in at least two first target nodes;
if the weight difference does not meet the threshold condition, adjusting one or more of the weights of the at least two first target nodes, the weights of the second target nodes and the edge weights of the at least two first target nodes and the second target nodes based on weight adjusting parameters, and re-determining the first target nodes matched with the second target nodes based on the adjusted weights.
8. The method of claim 7, wherein the weight adjustment parameter is a preset fixed value; alternatively, the first and second electrodes may be,
the weight adjustment parameter is the minimum of the weights of at least two of the first target nodes relative to the second target node.
9. The method of claim 1, further comprising, prior to generating an object association graph based on the object data for the first object and the object data for the second object:
setting a correlation range based on the first target, and determining the first target and the second target in the correlation range;
correspondingly, the generating a target association map based on the target data of the first target and the target data of the second target includes:
and for each association range, generating a target association graph based on the target data of the first target and the target data of the second target in the association range.
10. The method of claim 9, further comprising, after determining the first object and the second object within the relevance range:
for each association range, determining a first target number and a second target number in the association range;
if the first target number is greater than one and/or the second target number is greater than one, continuing to execute a step of generating a target association graph based on the target number of the first target and the target number of the second target in the association range;
if the first target number is less than one, or the second target number is less than one, determining the second target or the first target as a detection target;
and if the number of the first targets is equal to one and the number of the second targets is equal to one, performing association processing on the first targets and the second targets based on the association similarity of the first targets and the second targets to determine detection targets.
11. The method according to claim 10, wherein the associating the first target and the second target based on the associated similarity of the first target and the second target to determine the detection target comprises:
determining motion similarity and/or morphological similarity of the first target and the second target based on the target data of the first target and the target data of the second target;
and judging the motion similarity and/or the form similarity of the first target and the second target based on the association threshold, and determining the association relation of the first target and the second target.
12. An object associating apparatus, comprising:
the target identification module is used for acquiring first data acquired by a first type of sensor in a vehicle, identifying to obtain a first target based on the first data, acquiring second data acquired by a second type of sensor in the vehicle, and identifying to obtain a second target based on the second data;
an association graph generation module, configured to generate a target association graph based on target data of the first target and target data of the second target, where the target association graph includes target nodes corresponding to the first target and the second target, and edges between the target nodes, and the target nodes and the edges are provided with weights, respectively;
and the association processing module is used for performing association processing on the first target and the second target based on the weight corresponding to each target node and each edge in the target association graph to obtain the detection target of the vehicle.
13. An electronic device, characterized in that the electronic device comprises:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the object association method of any one of claims 1-11.
14. A computer-readable storage medium storing computer instructions for causing a processor to perform the object association method of any one of claims 1-11 when executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310139496.XA CN115827925A (en) | 2023-02-21 | 2023-02-21 | Target association method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310139496.XA CN115827925A (en) | 2023-02-21 | 2023-02-21 | Target association method and device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115827925A true CN115827925A (en) | 2023-03-21 |
Family
ID=85521956
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310139496.XA Pending CN115827925A (en) | 2023-02-21 | 2023-02-21 | Target association method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115827925A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107784042A (en) * | 2016-08-31 | 2018-03-09 | 高德软件有限公司 | A kind of object matching method and device |
| CN109556615A (en) * | 2018-10-10 | 2019-04-02 | 吉林大学 | The driving map generation method of Multi-sensor Fusion cognition based on automatic Pilot |
| CN110674719A (en) * | 2019-09-18 | 2020-01-10 | 北京市商汤科技开发有限公司 | Target object matching method and device, electronic equipment and storage medium |
| CN110927742A (en) * | 2019-11-19 | 2020-03-27 | 杭州飞步科技有限公司 | Obstacle tracking method, device, equipment and storage medium |
| CN114089329A (en) * | 2021-11-18 | 2022-02-25 | 重庆邮电大学 | Target detection method based on fusion of long and short focus cameras and millimeter wave radar |
-
2023
- 2023-02-21 CN CN202310139496.XA patent/CN115827925A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107784042A (en) * | 2016-08-31 | 2018-03-09 | 高德软件有限公司 | A kind of object matching method and device |
| CN109556615A (en) * | 2018-10-10 | 2019-04-02 | 吉林大学 | The driving map generation method of Multi-sensor Fusion cognition based on automatic Pilot |
| CN110674719A (en) * | 2019-09-18 | 2020-01-10 | 北京市商汤科技开发有限公司 | Target object matching method and device, electronic equipment and storage medium |
| CN110927742A (en) * | 2019-11-19 | 2020-03-27 | 杭州飞步科技有限公司 | Obstacle tracking method, device, equipment and storage medium |
| CN114089329A (en) * | 2021-11-18 | 2022-02-25 | 重庆邮电大学 | Target detection method based on fusion of long and short focus cameras and millimeter wave radar |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP4044117A1 (en) | Target tracking method and apparatus, electronic device, and computer-readable storage medium | |
| CN111783905B (en) | Target fusion method and device, storage medium and electronic equipment | |
| WO2022213632A1 (en) | Millimeter-wave radar calibration method and apparatus, and electronic device and roadside device | |
| WO2023155387A1 (en) | Multi-sensor target detection method and apparatus, electronic device and storage medium | |
| CN115471805B (en) | Point cloud processing and deep learning model training method and device and automatic driving vehicle | |
| CN114842445A (en) | Target detection method, device, equipment and medium based on multi-path fusion | |
| CN115797736A (en) | Method, device, equipment and medium for training target detection model and target detection | |
| CN115685249A (en) | Obstacle detection method and device, electronic equipment and storage medium | |
| CN115861959A (en) | Lane line identification method and device, electronic equipment and storage medium | |
| CN115909253A (en) | Target detection and model training method, device, equipment and storage medium | |
| CN115827925A (en) | Target association method and device, electronic equipment and storage medium | |
| CN115951344A (en) | Data fusion method and device for radar and camera, electronic equipment and storage medium | |
| CN116469073A (en) | Target identification method, device, electronic equipment, medium and automatic driving vehicle | |
| CN113920273B (en) | Image processing method, device, electronic equipment and storage medium | |
| CN112630736B (en) | Parameter determination method, device, equipment and storage medium of road side radar | |
| CN113902898A (en) | Training of target detection model, target detection method, device, equipment and medium | |
| CN114429631A (en) | Three-dimensional object detection method, device, equipment and storage medium | |
| CN114694138B (en) | Road surface detection method, device and equipment applied to intelligent driving | |
| CN117710459A (en) | Method, device and computer program product for determining three-dimensional information | |
| CN117392000B (en) | Noise removing method and device, electronic equipment and storage medium | |
| CN116363615B (en) | Data fusion method, device, vehicle and storage medium | |
| CN115346374B (en) | Intersection holographic perception method and device, edge computing equipment and storage medium | |
| EP4134843A2 (en) | Fusion and association method and apparatus for traffic objects in driving environment, and edge computing device | |
| CN117746386A (en) | Target object position sensing method, device and computer program product | |
| CN115661190A (en) | Target matching method, device, equipment and storage medium |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20230321 |
|
| RJ01 | Rejection of invention patent application after publication |