CN115359096A - Track prediction denoising method and device based on deep learning model and storage medium - Google Patents
Track prediction denoising method and device based on deep learning model and storage medium Download PDFInfo
- Publication number
- CN115359096A CN115359096A CN202211290414.3A CN202211290414A CN115359096A CN 115359096 A CN115359096 A CN 115359096A CN 202211290414 A CN202211290414 A CN 202211290414A CN 115359096 A CN115359096 A CN 115359096A
- Authority
- CN
- China
- Prior art keywords
- track
- curvature
- track points
- prediction
- deep learning
- 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.)
- Granted
Links
- 238000013136 deep learning model Methods 0.000 title claims abstract description 33
- 238000000034 method Methods 0.000 title claims abstract description 26
- 238000012549 training Methods 0.000 claims abstract description 103
- 238000004364 calculation method Methods 0.000 claims description 22
- 238000012216 screening Methods 0.000 claims description 12
- 238000001914 filtration Methods 0.000 claims description 8
- 238000004458 analytical method Methods 0.000 claims description 7
- 238000013075 data extraction Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 5
- 230000035772 mutation Effects 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004140 cleaning Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 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
- 238000005070 sampling Methods 0.000 description 1
- 238000012360 testing method 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
- G06T2207/30241—Trajectory
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Machines For Laying And Maintaining Railways (AREA)
Abstract
The invention discloses a track prediction denoising method, a track prediction denoising device and a storage medium based on a deep learning model, which comprise the steps of obtaining training samples, obtaining tracking track data, analyzing to obtain all track points, calculating Euclidean distances between a starting point and an end point of a tracking track, calculating the Euclidean distances between adjacent track points, summing the Euclidean distances, calculating a jitter value, and deleting the training samples with the jitter value larger than a jitter threshold value from a track prediction training sample set. The track prediction denoising method based on the deep learning model can effectively clean the training data and screen out high-quality training samples, thereby reducing model prediction problems caused by the quality problems of the training samples and improving the accuracy and stability of track prediction.
Description
Technical Field
The invention belongs to the field of unmanned vehicle trajectory prediction, and particularly relates to a trajectory prediction technology based on a deep learning model.
Background
The track prediction technology based on the deep learning model is the mainstream track prediction technology at present. The training samples of the track prediction model mainly comprise two types of representation methods, one is that data takes an image or point cloud as a training sample, and the other is that the track of a road traffic participant and map elements are vectorized, such as vector Net of a waymo behavior prediction algorithm, and are stored as the training samples through a graph data structure. Compared with the former, the vectorized data has stronger interpretability, is beneficial to constructing features, and has lower calculation cost for the model. Therefore, the main study here is a deep learning model using vectorized training samples.
For the trajectory prediction model, the training data is mainly composed of three parts: the method comprises the steps of a vehicle to be predicted (Agent), an obstacle around the vehicle to be predicted (Actor), and a map element (lane graph). For the prediction task, a historical x second track of the vehicle to be predicted and obstacles around the vehicle to be predicted and map elements around the vehicle to be predicted are generally selected as input, and a future y second track of the vehicle to be predicted is selected as a true value (GT, ground route) to train a neural network. In the training of the track prediction model, abnormal conditions such as circle hooking, reverse and the like can occur in the predicted track. The track can cause misjudgment of a downstream module, cause inching brake or take-over, and influence the performance of the unmanned vehicle. By manually sampling and screening samples, it is found that about 16% of training samples have true value jitter problem due to unstable perception or tracking, which makes the model fit wrong true value when training.
Disclosure of Invention
In order to solve the problem that the true value jitter of a training sample causes the fitting error true value of a track prediction model during training, the invention provides a method for denoising the training sample to improve the stability of the track prediction model.
In order to achieve the above object, an embodiment of the present invention provides a track prediction denoising method based on a deep learning model, which is characterized by including:
acquiring training samples from the trajectory prediction training sample set;
acquiring tracking track data from a training sample;
analyzing the tracking track data to obtain all track points corresponding to the track truth value;
calculating the Euclidean distance between the starting point and the end point of the tracking track;
traversing all track points, calculating Euclidean distances of adjacent track points and summing;
and calculating a jitter value according to the Euclidean distance between the starting point and the end point and the sum of the Euclidean distances of adjacent track points, and deleting the training samples of which the jitter value is greater than a jitter threshold value from the track prediction training sample set.
In order to achieve the above object, an embodiment of the present invention further provides a track prediction denoising device based on a deep learning model, which is characterized by including:
the training sample extraction module is used for acquiring training samples from the track prediction training sample set;
the tracking track data extraction module is used for acquiring tracking track data from the training sample;
the track analysis module is used for analyzing the tracking track data to obtain all track points corresponding to the track true value;
the first calculation module is used for calculating the Euclidean distance between the starting point and the end point of the tracking track;
the second calculation module is used for traversing all track points, calculating Euclidean distances of adjacent track points and summing the Euclidean distances;
the third calculation module is used for calculating a jitter value according to the sum of the Euclidean distance between the starting point and the end point and the Euclidean distance between adjacent track points;
the comparison module is used for comparing the jitter value with a jitter threshold value;
and the filtering module is used for deleting the training samples with the jitter values larger than the jitter threshold value from the track prediction training sample set.
In order to achieve the above object, an embodiment of the present invention further provides an apparatus, which includes a memory, a processor, and a program stored in the memory and executable on the processor, where the program implements the steps of the track prediction denoising method based on a deep learning model when executed by the processor.
In order to achieve the above object, an embodiment of the present invention further provides a storage medium, where the storage medium stores at least one program, and the at least one program is executable by at least one processor to implement the steps of the track prediction denoising method based on the deep learning model.
The invention has the beneficial effects that:
the track prediction denoising method based on the deep learning model can effectively clean the training data and screen out high-quality training samples, thereby reducing model prediction problems caused by the quality problems of the training samples and improving the accuracy and stability of track prediction.
Drawings
Fig. 1 is a flowchart of a track prediction denoising method based on a deep learning model in embodiment 1;
fig. 2 is a flowchart of a track prediction denoising method based on a deep learning model in embodiment 2;
fig. 3 is a block diagram of a track prediction denoising device based on a deep learning model in embodiment 3;
fig. 4 is a block diagram of a track prediction denoising device based on a deep learning model in embodiment 4;
fig. 5 is a block diagram showing the structure of the screening module in embodiment 4.
Detailed Description
Embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. It should be understood that the drawings and embodiments of the present application are for illustration purposes only and are not intended to limit the scope of the present application.
Example 1
The embodiment provides a track prediction denoising method based on a deep learning model, as shown in fig. 1, including:
Through an interface get _ agent _ IDs (), the tracking track IDs of all training samples in the database can be obtained, and the tracking track IDs correspond to the training samples one to one, so that the corresponding training samples are obtained.
By inputting the tracking track ID to the interface get _ track _ by _ ID (), the tracking track data corresponding to the training sample can be returned from the database.
And 103, analyzing all track points corresponding to the track truth value from the tracking track data of the training sample by using the conventional analysis method.
And 105, traversing all track points of the training sample, calculating Euclidean distances of every two adjacent track points, summing the Euclidean distances, and recording the Euclidean distances as perimeter.
And 106, calculating a jitter value score according to the sum of the Euclidean distance between the starting point and the end point and the Euclidean distance between adjacent track points, and deleting the training samples with jitter values larger than a jitter threshold value from the track prediction training sample set.
And 107, acquiring the next training sample, and repeating the steps 101-106 until all the training samples are traversed.
Wherein score = perimeter/traj _ len, for a smooth trajectory, the jitter value should be close to 1, and the jitter value is too large, which indicates that the predicted trajectory is abnormal, and the quality of the corresponding training sample is low. The setting of the jitter threshold can be set reasonably according to the expected target of the trajectory tracking. The training samples with jitter values larger than the jitter threshold value are deleted from the track prediction training sample set, namely the training samples with low quality and track true value jitter are deleted, and the training sample set is subjected to noise removal and cleaning, so that a model cannot be subjected to false true value fitting during training, model prediction problems caused by the quality problems of the training samples are reduced, and the accuracy and the stability of track prediction are improved.
The track prediction denoising method based on the deep learning model is mainly applied to a behavior or track prediction module in an automatic driving technology, and a prediction model developed based on the deep learning technology on the front edge often needs a specific form of input. Particularly in the model training stage, correctly available training and testing samples can be obtained for dynamically and continuously expanded self-collected data, and the predicted trajectory quality of the prediction model can be better improved from the data side.
Example 2
The present embodiment provides a trajectory prediction denoising method based on a deep learning model, including steps 101 to 107 of embodiment 1, as shown in fig. 2, after cleaning a training sample in step 107 and deleting the training sample having a trajectory true value and a trajectory jitter value, the method further includes:
and step 108, acquiring tracking track data for each residual training sample in the track prediction training sample set, analyzing to obtain all track points corresponding to the track true value, traversing all track points, and calculating the curvature of each track point and the curvature variation of adjacent track points.
And the curvature calculation of each track point determines a circle by taking every adjacent 3 track points as a group for calculation, and the curvature difference of every two adjacent track points is subtracted to obtain an absolute value so as to obtain the curvature variation.
And step 109, screening out training samples with track point curvature absolute values larger than a curvature threshold value or with abrupt changes in track point curvature.
Step 109 specifically includes: and setting a curvature threshold and a curvature variation threshold, comparing the curvature absolute value of the track point with the curvature threshold, comparing the curvature variation of the adjacent track point with the curvature variation threshold, and judging that the curvature of the track point has a sudden change when the curvature variation of the adjacent track point is greater than the curvature variation threshold. When the curvature absolute value of a certain track point is larger than the curvature threshold, or the curvature variation of two adjacent track points is larger than the curvature variation threshold, namely, the curvature of the track point has sudden change, the prediction effect of the track prediction model can be influenced. In the embodiment, the training samples with the screened true track value and the overlarge curvature or mutation are put into the training sample set to be repaired for repair, so that the model prediction problem caused by the quality problem of the training samples can be further reduced, and the accuracy and the stability of track prediction are improved. The curvature threshold and the curvature change amount threshold can be reasonably set according to the expected target of the track tracking.
Example 3
The embodiment provides a track prediction denoising device based on a deep learning model, as shown in fig. 3, including a training sample extraction module 1, a tracking track data extraction module 2, a track analysis module 3, a first calculation module 4, a second calculation module 5, a third calculation module 6, a comparison module 7, and a filtering module 8. The training sample extraction module 1 obtains training samples from a track prediction training sample set and inputs the training samples into the tracking track data extraction module 2, the tracking track data extraction module 2 obtains tracking track data from the training samples and inputs the tracking track data into the track analysis module 3, and the track analysis module 3 analyzes the tracking track data by using the prior art to obtain all track points corresponding to a track true value. The first calculation module 4 reads the start point and end point data of the tracking track, and calculates the euclidean distance between the start point and the end point. The second calculation module 5 traverses all the trace points, calculates the euclidean distance between every two adjacent trace points and sums the calculated euclidean distances. The third calculation module 6 inputs the euclidean distance between the starting point and the end point and the sum of the euclidean distances between adjacent track points, and calculates the jitter value by using the following formula: score = perimeter/traj _ len, where score represents the jitter value, traj _ len represents the euclidean distance of the starting point from the end point, and perimeter represents the sum of the euclidean distances of adjacent track points. The comparison module 7 inputs the calculated jitter value, compares the calculated jitter value with a jitter threshold value, and inputs the comparison result into the filtering module 8. When the true track jitter value of a training sample is greater than the jitter threshold, the filtering module 8 obtains the corresponding training sample according to the tracking track ID of the training sample, and deletes the training sample whose jitter value is greater than the jitter threshold from the track prediction training sample set.
Example 4
The embodiment provides a track prediction denoising device based on a deep learning model, which includes a training sample extraction module 1, a tracking track data extraction module 2, a track analysis module 3, a first calculation module 4, a second calculation module 5, a third calculation module 6, a comparison module 7, and a filtering module 8 in embodiment 3, as shown in fig. 4, and further includes a fourth calculation module 9, a fifth calculation module 10, and a screening module 11. After the training samples with shaking track truth values are cleaned by the filtering module 8, the multiplexing training sample extracting module 1 of the embodiment intensively obtains training samples from the rest training samples, the multiplexing tracking track data extracting module 2 obtains tracking track data of the training samples, the multiplexing track analyzing module 3 analyzes the training samples to obtain all track points corresponding to the track truth values, then the fourth calculating module traverses all track points and calculates curvatures of the track points, the fifth calculating module 10 traverses all track points and subtracts the curvatures of adjacent track points to obtain absolute values to obtain curvature variation of the adjacent track points, and finally the screening module 11 screens out the training samples with overlarge track point curvature absolute values or with sudden changes in track point curvature, and puts the training samples into a training sample set to be repaired for repair, so that the model prediction problem caused by the self quality problem of the training samples can be further reduced, and the accuracy and the stability of track prediction are improved.
As shown in fig. 5, the screening module 11 includes a first comparing unit 1101 for comparing the curvature absolute value of the trace point with a curvature threshold; the second comparing unit 1102 is configured to compare the curvature variation of the adjacent track points with a curvature variation threshold, and determine that the curvature of the track points has a sudden change when the curvature variation of the adjacent track points is greater than the curvature variation threshold; the first comparison unit 1101 and the second comparison unit 1102 output comparison results to the screening unit 1103, the screening unit 1103 extracts training samples with track point curvature absolute values larger than a curvature threshold value or with abrupt changes in track point curvature, and the training samples are put into a training sample set to be repaired for repair.
Example 5
The present embodiment provides an apparatus, which includes a memory, a processor, and a program stored on the memory and executable on the processor, and when the program is executed by the processor, the program implements the steps of the trajectory prediction denoising method based on the deep learning model in the foregoing embodiments.
Example 6
The present embodiment provides a storage medium storing at least one program, which is executable by at least one processor, to implement the steps of the trajectory prediction denoising method based on the deep learning model in the above embodiments.
Those of ordinary skill in the art will understand that: although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art will recognize that changes may be made in the form and details of the embodiments without departing from the spirit and scope of the invention. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (10)
1. A track prediction denoising method based on a deep learning model is characterized by comprising the following steps:
acquiring training samples from the trajectory prediction training sample set;
acquiring tracking track data from a training sample;
analyzing the tracking track data to obtain all track points corresponding to the track truth value;
calculating the Euclidean distance between the starting point and the end point of the tracking track;
traversing all track points, calculating Euclidean distances of adjacent track points and summing;
and calculating a jitter value according to the Euclidean distance between the starting point and the end point and the sum of the Euclidean distances of adjacent track points, and deleting the training samples of which the jitter value is greater than a jitter threshold value from the track prediction training sample set.
2. The method for denoising trajectory prediction based on deep learning model as claimed in claim 1, further comprising:
residual training samples in the track prediction training sample set are analyzed to obtain tracking track data, all track points corresponding to the track true value are obtained through analysis, all track points are traversed, and the curvature of the track points and the curvature variation of adjacent track points are calculated;
and screening out training samples with the track point curvature absolute value larger than a curvature threshold value or with abrupt changes in the track point curvature.
3. The method for denoising trajectory prediction based on deep learning model as claimed in claim 2, comprising:
subtracting the curvatures of adjacent track points to obtain an absolute value to obtain the curvature variation of the adjacent track points;
and comparing the curvature variation of the adjacent track points with a curvature variation threshold, and judging that the curvature of the track points has a sudden change when the curvature variation of the adjacent track points is greater than the curvature variation threshold.
4. The deep learning model-based trajectory prediction denoising method of claim 1, wherein the calculating of the jitter value comprises: the formula score = perimeter/traj _ len is adopted, where score represents the jitter value, traj _ len represents the euclidean distance of the starting point and the end point, and perimeter represents the sum of the euclidean distances of adjacent track points.
5. A track prediction denoising device based on a deep learning model is characterized by comprising:
the training sample extraction module is used for acquiring training samples from the track prediction training sample set;
the tracking track data extraction module is used for acquiring tracking track data from the training sample;
the track analyzing module is used for analyzing the tracking track data to obtain all track points corresponding to the track truth value;
the first calculation module is used for calculating the Euclidean distance between the starting point and the end point of the tracking track;
the second calculation module is used for traversing all track points, calculating Euclidean distances of adjacent track points and summing the Euclidean distances;
the third calculation module is used for calculating a jitter value according to the sum of the Euclidean distance between the starting point and the end point and the Euclidean distance between adjacent track points;
the comparison module is used for comparing the jitter value with a jitter threshold value;
and the filtering module is used for deleting the training samples with the jitter values larger than the jitter threshold value from the track prediction training sample set.
6. The deep learning model-based trajectory prediction denoising apparatus of claim 5, further comprising:
the fourth calculation module is used for traversing all track points and calculating the curvatures of the track points;
the fifth calculation module is used for traversing all track points, and subtracting the curvatures of the adjacent track points to obtain an absolute value to obtain the curvature variation of the adjacent track points;
and the screening module is used for screening out the training samples with the track point curvature absolute value larger than a curvature threshold value or the track point curvature with mutation.
7. The deep learning model-based trajectory prediction denoising apparatus of claim 6, wherein the filtering module comprises:
the first comparison unit is used for comparing the curvature absolute value of the track point with a curvature threshold value;
the second comparison unit is used for comparing the curvature variation of the adjacent track points with a curvature variation threshold, and when the curvature variation of the adjacent track points is larger than the curvature variation threshold, the curvature of the track points is judged to have sudden change;
and the screening unit is used for extracting the training samples of which the curvature absolute values of the track points are greater than the curvature threshold value or the curvatures of the track points have sudden changes.
8. The apparatus for denoising trajectory prediction based on deep learning model as claimed in claim 5, wherein the third calculating module calculating the jitter value comprises:
the formula score = perimeter/traj _ len is adopted, where score represents the jitter value, traj _ len represents the euclidean distance of the starting point and the end point, and perimeter represents the sum of the euclidean distances of adjacent track points.
9. An apparatus comprising a memory, a processor, a program stored on the memory and executable on the processor, wherein the program when executed by the processor implements the steps of the deep learning model based trajectory prediction denoising method of any one of claims 1-4.
10. A storage medium storing at least one program executable by at least one processor to perform the steps of the deep learning model-based trajectory prediction denoising method of any one of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211290414.3A CN115359096B (en) | 2022-10-21 | 2022-10-21 | Track prediction denoising method and device based on deep learning model and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211290414.3A CN115359096B (en) | 2022-10-21 | 2022-10-21 | Track prediction denoising method and device based on deep learning model and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115359096A true CN115359096A (en) | 2022-11-18 |
| CN115359096B CN115359096B (en) | 2023-05-09 |
Family
ID=84008931
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211290414.3A Active CN115359096B (en) | 2022-10-21 | 2022-10-21 | Track prediction denoising method and device based on deep learning model and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115359096B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112732857A (en) * | 2021-01-20 | 2021-04-30 | 腾讯科技（深圳）有限公司 | Road network processing method, road network processing device, electronic equipment and storage medium |
| CN113283653A (en) * | 2021-05-27 | 2021-08-20 | 大连海事大学 | Ship track prediction method based on machine learning and AIS data |
| CN113793031A (en) * | 2021-09-15 | 2021-12-14 | 中海油安全技术服务有限公司 | Submarine pipeline risk prediction method and device |
| CN114802303A (en) * | 2022-04-26 | 2022-07-29 | 中国第一汽车股份有限公司 | Obstacle trajectory prediction method, obstacle trajectory prediction device, electronic device, and storage medium |
| CN114987546A (en) * | 2022-06-10 | 2022-09-02 | 中国第一汽车股份有限公司 | Training method, device and equipment of trajectory prediction model and storage medium |
| CN115002679A (en) * | 2022-07-18 | 2022-09-02 | 北京航天泰坦科技股份有限公司 | Trajectory deviation rectifying processing method and device |
| CN115186576A (en) * | 2022-06-17 | 2022-10-14 | 同济大学 | Non-motor vehicle track prediction method |
-
2022
- 2022-10-21 CN CN202211290414.3A patent/CN115359096B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112732857A (en) * | 2021-01-20 | 2021-04-30 | 腾讯科技（深圳）有限公司 | Road network processing method, road network processing device, electronic equipment and storage medium |
| CN113283653A (en) * | 2021-05-27 | 2021-08-20 | 大连海事大学 | Ship track prediction method based on machine learning and AIS data |
| CN113793031A (en) * | 2021-09-15 | 2021-12-14 | 中海油安全技术服务有限公司 | Submarine pipeline risk prediction method and device |
| CN114802303A (en) * | 2022-04-26 | 2022-07-29 | 中国第一汽车股份有限公司 | Obstacle trajectory prediction method, obstacle trajectory prediction device, electronic device, and storage medium |
| CN114987546A (en) * | 2022-06-10 | 2022-09-02 | 中国第一汽车股份有限公司 | Training method, device and equipment of trajectory prediction model and storage medium |
| CN115186576A (en) * | 2022-06-17 | 2022-10-14 | 同济大学 | Non-motor vehicle track prediction method |
| CN115002679A (en) * | 2022-07-18 | 2022-09-02 | 北京航天泰坦科技股份有限公司 | Trajectory deviation rectifying processing method and device |
Non-Patent Citations (1)
| Title |
|---|
| 林毅等: "基于GMM的航班轨迹预测算法研究", 《工程科学与技术》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115359096B (en) | 2023-05-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110084095B (en) | Lane line detection method, lane line detection apparatus, and computer storage medium | |
| CN110348445B (en) | Instance segmentation method fusing void convolution and edge information | |
| CN109213755B (en) | A kind of traffic flow data cleaning and restorative procedure based on Time-space serial | |
| KR101067437B1 (en) | Lane detection method and Detecting system using the same | |
| JP2007510993A (en) | Object detection in images | |
| CN115222650A (en) | Mixed industrial part defect detection algorithm | |
| CN113780110A (en) | Method and device for detecting weak and small targets in image sequence in real time | |
| CN115359096A (en) | Track prediction denoising method and device based on deep learning model and storage medium | |
| CN111862088B (en) | Method for rapidly calculating crack width | |
| CN115481144A (en) | Method and device for correcting track prediction model training samples and storage medium | |
| CN116596895A (en) | Substation equipment image defect identification method and system | |
| CN116152758A (en) | Intelligent real-time accident detection and vehicle tracking method | |
| CN105279761A (en) | A background modeling method based on sample local density outlier detection | |
| CN109934193A (en) | Prior-constrained anti-of global context blocks face super-resolution method and its system | |
| CN113673383B (en) | Time-space domain obstacle detection method and system for complex road scene | |
| CN104992436A (en) | Image segmentation method for natural scene | |
| CN110705545B (en) | Obstacle identification method and device, storage medium and inspection robot | |
| KR102157007B1 (en) | Self-Learning Device and Device for Object Detection | |
| CN114863411A (en) | License plate recognition method and device | |
| CN113487620A (en) | Railway insulation section detection method and device | |
| JP2995650B2 (en) | Ruled line removal method | |
| CN116612451B (en) | Road edge identification method, device and equipment for unmanned sweeper and storage medium | |
| CN117114197B (en) | Symptom migration range identification method based on deep learning | |
| CN104573692B (en) | License plate binarization method based on fuzzy degradation model | |
| Liu et al. | A fast background update mechanism for vehicle detection in urban roads |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20230426 Address after: 215100 station 601-b11, Tiancheng information building, No. 88, nantiancheng Road, Xiangcheng District, Suzhou City, Jiangsu Province Applicant after: Zhongzhixing (Suzhou) Technology Co.,Ltd. Applicant after: Tianyi Transportation Technology Co.,Ltd. Address before: 215100 station 601-b11, Tiancheng information building, No. 88, nantiancheng Road, Xiangcheng District, Suzhou City, Jiangsu Province Applicant before: Zhongzhixing (Suzhou) Technology Co.,Ltd. |