CN112766538A - Decision modeling method for pedestrian crossing road based on SVM algorithm - Google Patents
Decision modeling method for pedestrian crossing road based on SVM algorithm Download PDFInfo
- Publication number
- CN112766538A CN112766538A CN202011570760.8A CN202011570760A CN112766538A CN 112766538 A CN112766538 A CN 112766538A CN 202011570760 A CN202011570760 A CN 202011570760A CN 112766538 A CN112766538 A CN 112766538A
- Authority
- CN
- China
- Prior art keywords
- pedestrian
- decision
- data set
- data
- svm algorithm
- 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 35
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 33
- 238000012360 testing method Methods 0.000 claims abstract description 39
- 230000001815 facial effect Effects 0.000 claims abstract description 21
- 230000001419 dependent effect Effects 0.000 claims abstract description 5
- 238000012549 training Methods 0.000 claims description 17
- 238000012545 processing Methods 0.000 claims description 4
- 238000011161 development Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 3
- 238000005290 field theory Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 210000003734 kidney Anatomy 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Strategic Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- General Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Development Economics (AREA)
- Human Computer Interaction (AREA)
- Game Theory and Decision Science (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention relates to the technical field of traffic safety, and particularly discloses a pedestrian crossing road decision modeling method based on an SVM algorithm, which comprises the following steps: acquiring actual pedestrian data, and acquiring facial feature data of a plurality of actual pedestrians crossing a road by using face recognition equipment; extracting the viewing direction switching times and the duration of keeping a side-looking state of a pedestrian who violates the crosswalk within a period of time to obtain an original data set; dividing an original data set into a passing data set and a waiting data set, and respectively adding a passing decision type label and a waiting decision type label; when a pedestrian crossing road decision model based on an SVM algorithm is trained, the number of times of switching the viewing direction of the pedestrian and the duration of keeping a side-looking state are used as independent variables, and a decision type label is used as a dependent variable; and (6) testing the model. The invention has the advantages of low modeling cost, convenient data acquisition, small calculated amount, high operation speed and accurate prediction.
Description
Technical Field
The invention relates to the technical field of traffic safety, in particular to a pedestrian crossing road decision modeling method based on an SVM algorithm.
Background
In real life, the pedestrian violates crossing of the road, including crossing of the road at an intersection, a pedestrian road and a traffic light. The illegal passing of the pedestrian poses a non-negligible threat to the running vehicle and the pedestrian itself. With the development of driving assistance techniques, more and more driving assistance techniques are gradually reducing the driving burden of the driver. The rapid development of image recognition technology, especially the development of face recognition technology and computer computing power, makes it possible to predict the decision-making intention of pedestrians through the face recognition technology. In order to solve the decision-making purpose of accurately distinguishing the crossing of the pedestrian on the road, increase the function of a vehicle auxiliary driving system and ensure the safety of the pedestrian crossing the road when the vehicle runs through the intersection, an accurate pedestrian crossing road decision model is required to be established. In the prior art, an accurate pedestrian crossing road decision model and a modeling method thereof are not disclosed.
The Chinese patent with the publication number of CN109871738A and the name of 'a pedestrian movement intention identification method adaptive to a self-mixed environment' provides a method for constructing a pedestrian movement intention identification model based on a phase field theory and a fuzzy logic method. The method mainly uses field theory and fuzzy logic, and does not relate to the method for carrying out relevant model training based on SVM algorithm.
The chinese patent with the application number of CN201910066829.4 and the name of "a system and a method for detecting street crossing intention of non-motor vehicles or pedestrians" proposes a method for determining the street crossing intention of pedestrians or motor vehicles according to the transverse position data of non-motor vehicles, the number of times of looking back by riders or the number of times of looking back by turns of pedestrians on road sections, and the transverse position and the moving speed of pedestrians on sidewalks, but does not relate to a method for modeling according to data acquired when a plurality of actual pedestrians cross a street, and the accuracy of predicting the street crossing intention of pedestrians is not reliable.
Therefore, a modeling method with small calculation amount, low development cost and high accuracy for pedestrian crossing road decision models needs to be developed.
Disclosure of Invention
In order to solve the problems in the prior art, the main object of the present invention is to provide a modeling method for pedestrian crossing road decision, which establishes a pedestrian crossing road decision model based on SVM algorithm according to collected facial feature data of actual pedestrians crossing the road.
In order to achieve the aim, the invention provides a pedestrian crossing road decision modeling method based on an SVM algorithm, which comprises the following steps:
step one, collecting actual pedestrian data: at an intersection comprising a sidewalk and a traffic signal lamp, acquiring facial feature data of a plurality of actual pedestrians crossing a road by using face recognition equipment;
step two, processing test data: the method specifically comprises the following substeps:
1) removing the facial feature data of the pedestrians outside the range L of outward offset of the outer contour line of the sidewalk;
2) calculating and extracting the number of times of viewing direction switching and the duration of keeping a side-looking state of a pedestrian crossing a sidewalk in violation of a traffic signal lamp when the pedestrian passes through a certain period of time to obtain an original data set; the switching times of the viewing directions of the pedestrians and the duration of keeping the side-looking state are determined according to facial feature data of the pedestrians in a period of time when the pedestrians start to walk;
3) dividing the original data set into a passing data set and a waiting data set according to whether the pedestrian passes through the sidewalk in a violation mode, adding a passing decision type label into data of the passing data set, and adding a waiting decision type label into data of the waiting data set;
step three, training a model: training by utilizing the traversing data set, the waiting data set and the SVM algorithm to obtain a pedestrian crossing road decision model based on the SVM algorithm; when the model is trained, the number of times of switching the viewing direction of the pedestrian and the duration of keeping the side-looking state are used as independent variables, and a pedestrian decision type label is used as a dependent variable;
and step four, testing the model.
Further, the facial feature data comprise a pedestrian face yaw angle and a face pitch angle; when the pitch angle of the face is larger than a certain threshold value gamma, the pedestrian is considered to be looking over;
when the leftward yaw angle of the face is larger than a certain threshold value theta and the pitch angle of the face is larger than a certain threshold value gamma, the pedestrian is considered to be looking leftward;
when the right yaw angle of the face is larger than a certain threshold value theta and the pitch angle of the face is larger than a certain threshold value gamma, the pedestrian is considered to be viewed rightwards;
when the leftward yaw angle of the face is smaller than a certain threshold value theta, the rightward yaw angle of the face is smaller than a certain threshold value theta, and the pitch angle of the face is larger than a certain threshold value gamma, the pedestrian is considered to be looking forward;
the viewing direction switching times refer to the switching times of the pedestrian viewing direction from one direction to the other direction;
the length of time for which the side-viewing state is maintained indicates a length of time for which the pedestrian remains in the leftward viewing state or in the rightward viewing state.
In a preferred embodiment, the threshold γ is set to 60 °, and the threshold θ is set to 30 °.
In a preferred embodiment, in the sub-step 1) of the second step, the outward offset of the outer contour line of the sidewalk is set to 1 m.
Further, in the substep 2) of the second step, the traffic signal lamp is a traffic light, and when the red light is in a bright state, the pedestrian crossing the sidewalk is judged as the illegal pedestrian crossing the sidewalk.
In a specific embodiment, in sub-step 3) of the second step, the walk through decision type tag is set to a number "1", and the wait decision type tag is set to a number "0".
Further, in the third step, before training the decision model of the pedestrian crossing the road based on the SVM algorithm, data of pedestrians in a certain proportion R in the crossing data set and data of pedestrians in the same proportion R in the waiting data set are randomly selected as training data sets, and other data are selected as test data sets.
In a preferred embodiment, the ratio R is set to 80%.
Further, during model testing, the number of times of view direction switching of a pedestrian of a certain test data point in a test data set and the duration of keeping a side-looking state are used as input variables, a predicted pedestrian decision type label is calculated and output through the SVM algorithm-based pedestrian crossing road decision model, if the predicted pedestrian decision type label is the same as an actual decision type label, the model testing is successful for the test data point, and if not, the model testing fails.
If the ratio of the number of the successfully tested test data points to the total number of the data points in the test data set exceeds a preset threshold value P, the model precision is acceptable, and the modeling is successful. Otherwise, more test data needs to be collected for model training.
In a preferred embodiment, the predetermined threshold P is set to 85%. The threshold value P is set according to the model accuracy requirement, and the larger the threshold value P is, the higher the model accuracy is.
Due to the adoption of the technical scheme, the invention has the following beneficial effects: according to the method, the face characteristic data of the actual pedestrian crossing the road are collected, the number of times of switching the viewing direction of the pedestrian and the duration of keeping the side-looking state are used as independent variables, the pedestrian decision type label is used as a dependent variable, and a pedestrian crossing road decision model is obtained based on SVM training. The pedestrian crossing road decision model based on the SVM algorithm can accurately predict the decision intention of the pedestrian and overcome the defects of the prior art.
Drawings
FIG. 1 is a schematic flow chart of the steps of a pedestrian crossing road decision modeling method based on SVM algorithm according to the present invention.
Fig. 2 is a schematic diagram of a pedestrian detection area after the outer contour line of the pedestrian path is shifted outward by a certain distance L.
Fig. 3 is a coordinate system established when the face recognition device calculates the face feature data of the pedestrian.
Detailed Description
In order to make the technical solution of the embodiments of the present invention better understood, the technical solution of the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments, 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 equivalent changes and modifications by one skilled in the art based on the embodiments of the present invention, shall fall within the scope of the present invention.
Referring to fig. 1, the present embodiment provides a modeling method for pedestrian crossing road decision based on SVM, which includes the following steps:
s1, collecting actual pedestrian data by using face recognition equipment:
the places where the pedestrian data collection is performed include, but are not limited to, intersections including sidewalks and traffic lights (traffic lights in this embodiment), and facial feature data of a plurality of actual pedestrians crossing the road are collected through preset face recognition equipment, where the facial feature data include, but are not limited to, facial yaw angles and facial pitch angles of the pedestrians. In this embodiment, face identification equipment can adopt visible light, near-infrared people face to gather in face identification equipment of an organic whole, such as commercially available china kidney science and technology AI intelligence binocular camera, and the model is AI1901, installs it on the vehicle, can be used to carry out face identification to the pedestrian. The face recognition device is the prior art, and the specific structure and working principle thereof are known to those skilled in the art, and are not described herein.
S2, processing test data, and dividing training data and test data:
processing the test data includes:
1) and removing the facial feature data of the pedestrians beyond the range L of a certain distance of outward deviation of the outer contour line of the sidewalk. As shown in fig. 2, the present invention is concerned with pedestrians in a pedestrian detection area whose outer contour line of the sidewalk is shifted outward by a certain distance L, and recognizes the outer contour line a1 of the sidewalk, and obtains a pedestrian detection area a2 after shifting outward by L.
2) According to whether the pedestrian runs across a traffic signal lamp or not, calculating and extracting the number of times of viewing direction switching and the duration of keeping a side-looking state of the pedestrian which runs across the sidewalk in violation in a period of time t being 5s to obtain an original data set; wherein, the times of view direction switching of the pedestrian and the duration of keeping the side-looking state are determined according to the facial feature data of the pedestrian in a period of starting to walk. In this embodiment, the identification of the pedestrian path outer contour line and the traffic signal lamp is realized by the AI intelligent binocular camera.
When the face pitch angle is greater than a certain threshold γ, the pedestrian is considered to be looking.
When the leftward yaw angle of the face is greater than a certain threshold value theta, and the pitch angle of the face is greater than a certain threshold value gamma, the face is considered to be viewed leftward; when the yaw angle to the right of the face is greater than a certain threshold value theta, and the pitch angle of the face is greater than a certain threshold value gamma, 60 deg., the pedestrian is considered to be looking to the right.
When the face leftward yaw angle is smaller than a certain threshold value theta, the face rightward yaw angle is smaller than a certain threshold value theta, 30 degrees, and the face pitch angle is larger than a certain threshold value gamma, 60 degrees, the pedestrian is considered to be looking forward.
The viewing direction switching times refer to the switching times of the viewing directions of the pedestrians, for example, the pedestrians change from viewing to viewing forwards from viewing leftwards, and the viewing directions can be regarded as one-time switching of the viewing directions.
The length of time for which the side-viewing state is maintained indicates a length of time for which the pedestrian remains in the leftward viewing state or in the rightward viewing state.
As shown in fig. 3, it represents a coordinate system established when the face recognition device calculates the facial feature data of the pedestrian, when the face is right ahead, the center of the head of the human is taken as the origin, the front-back direction is an x-axis, the left-right direction is a y-axis, the up-down direction is a z-axis, the face yaw angle is the angle between the x-axis and the y-axis when the face swings left and right, and the face pitch angle represents the angle between the x-axis and the z-axis when the face looks up or down.
The data dividing step includes:
3) dividing the original data set into a passing data set and a waiting data set according to whether the pedestrian passes through the sidewalk in a violation mode, adding a passing decision type label into the data of the passing data set, and adding a waiting decision type label into the data of the waiting data set.
In this embodiment, a decision type label "1" is added to the traversal data set to indicate that the related data belongs to the traversal decision class. And adding a decision type label of '0' in the waiting data set to indicate that the related data belongs to the waiting decision class.
4) And randomly selecting test data of a certain proportion of pedestrians (R is 80%) in the passing data set and test data of the same proportion of pedestrians in the waiting data set as training data sets, and using other data as testing data sets.
S3, training by using training data and an SVM algorithm to obtain a pedestrian decision model based on the SVM:
based on a training data set and an SVM algorithm, when a pedestrian based on the SVM algorithm traverses a road decision model, the times of switching the viewing direction of the pedestrian and the duration of keeping a side-looking state are used as independent variables, and a pedestrian decision type label is used as a dependent variable.
S4, testing the pedestrian decision model based on the SVM by using the test data:
when the pedestrian crossing the road decision model based on the SVM is tested, the pedestrian checking direction switching times and the side-looking state keeping duration of the test data points in the test data set are used as model input, and the predicted pedestrian decision type label is obtained through model calculation and output. If the predicted decision type label is the same as the actual decision type label, the decision model is successfully predicted for the test data point, otherwise, the prediction fails.
In this embodiment, the number of successfully tested test data points accounts for 88.3% of the total data amount of the test data set, that is, the ratio of the number of successfully tested pedestrians to the total pedestrian amount in the total test data set is 88.3%, and if the predetermined threshold P is exceeded, it is determined that the accuracy of the pedestrian crossing the road decision model trained by the above modeling method is acceptable, and the modeling is successful. Otherwise, more test data needs to be collected for model training.
When the pedestrian crossing road decision model is applied, according to a traffic signal lamp and when a red light is turned on, facial feature data of pedestrians with outer contour lines of a sidewalk outwards deviating within a certain range L & lt1 m & gt, namely facial yaw angles and pitch angles are collected, the number of times of switching of the viewing direction and the duration of keeping a side-looking state of the pedestrians within a period of time t & lt5 s & gt are calculated, and according to the established pedestrian crossing road decision model based on an SVM algorithm, the decision type of the pedestrians which violate the rules and pass through the sidewalk can be accurately predicted, namely the crossing decision or the waiting decision.
When the decision type of the model for predicting the pedestrian is a passing decision, namely the system identifies that the pedestrian crosses the road, the system sends a prompt that the pedestrian crosses the road to the driver, so that the vehicle can decelerate in time to avoid the pedestrian, and the traffic safety of the vehicle and the surrounding pedestrians is improved.
Alternatively, when the system recognizes that a pedestrian crosses the road, it notifies a vehicle within 50m of the surroundings to pay attention to the driving safety. The computing system is deployed at the cloud, image information of the relevant cameras is collected through the Internet of things system, and after pedestrians crossing the road are processed and identified, the 'pedestrians in violation' noticing message is sent to vehicles within 50n of the periphery, and the messages are displayed on screens of the vehicles or informed in a voice mode.
Compared with the prior art, the modeling method has the advantages that the cost is low, the data acquisition is convenient, the modeling is carried out by using the SVM algorithm, the calculated amount of the obtained model is small, the operation speed is high, and the limitation of the prior art is overcome to a certain extent.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention; also, the above description should be understood as being readily apparent to those skilled in the relevant art and can be implemented, and therefore, other equivalent changes and modifications without departing from the concept disclosed herein are intended to be included within the scope of the present invention.
Claims (10)
1. A pedestrian crossing road decision modeling method based on SVM algorithm is characterized by comprising the following steps:
step one, collecting actual pedestrian data: at an intersection comprising a sidewalk and a traffic signal lamp, acquiring facial feature data of a plurality of actual pedestrians crossing a road by using face recognition equipment;
step two, processing test data: the method specifically comprises the following substeps:
1) removing the facial feature data of the pedestrians outside the range L of outward offset of the outer contour line of the sidewalk;
2) calculating and extracting the number of times of viewing direction switching and the duration of keeping a side-looking state of a pedestrian crossing a sidewalk in violation of a traffic signal lamp when the pedestrian passes through a certain period of time to obtain an original data set; the switching times of the viewing directions of the pedestrians and the duration of keeping the side-looking state are determined according to facial feature data of the pedestrians in a period of time when the pedestrians start to walk;
3) dividing the original data set into a passing data set and a waiting data set according to whether the pedestrian passes through the sidewalk in a violation mode, adding a passing decision type label into data of the passing data set, and adding a waiting decision type label into data of the waiting data set;
step three, training a model: training by utilizing the traversing data set, the waiting data set and the SVM algorithm to obtain a pedestrian crossing road decision model based on the SVM algorithm; when the model is trained, the number of times of switching the viewing direction of the pedestrian and the duration of keeping the side-looking state are used as independent variables, and a pedestrian decision type label is used as a dependent variable;
and step four, testing the model.
2. The modeling method for pedestrian cross-road decision of SVM algorithm of claim 1, wherein the facial feature data includes pedestrian facial yaw angle, facial pitch angle; when the pitch angle of the face is larger than a certain threshold value gamma, the pedestrian is considered to be looking over;
when the leftward yaw angle of the face is larger than a certain threshold value theta and the pitch angle of the face is larger than a certain threshold value gamma, the pedestrian is considered to be looking leftward;
when the right yaw angle of the face is larger than a certain threshold value theta and the pitch angle of the face is larger than a certain threshold value gamma, the pedestrian is considered to be viewed rightwards;
when the leftward yaw angle of the face is smaller than a certain threshold value theta, the rightward yaw angle of the face is smaller than a certain threshold value theta, and the pitch angle of the face is larger than a certain threshold value gamma, the pedestrian is considered to be looking forward;
the viewing direction switching times refer to the switching times of the pedestrian viewing direction from one direction to the other direction;
the length of time for which the side-viewing state is maintained indicates a length of time for which the pedestrian remains in the leftward viewing state or in the rightward viewing state.
3. A pedestrian crossing road decision modelling method of the SVM algorithm in accordance with claim 2 wherein said threshold γ is set at 60 ° and said threshold θ is set at 30 °.
4. A modeling method for pedestrian crossing road decision based on SVM algorithm as claimed in any one of claims 1-3 wherein in sub-step 1) of step two the outer contour line of the sidewalk is shifted outward by a certain distance L set to 1 m.
5. The modeling method for pedestrian crossing road decision based on SVM algorithm according to any of claims 1-3, characterized in that in the substep 2) of the second step, the traffic signal lamp is a traffic light, and when the red light is in a bright state, the pedestrian crossing the sidewalk is judged as illegal crossing the sidewalk.
6. The modeling method for pedestrian crossing road decision based on SVM algorithm according to any of claims 1-3, characterized in that in the third step, before training the decision model for pedestrian crossing road based on SVM algorithm, the data of pedestrian with a certain proportion R in the crossing data set and the data of pedestrian with the same proportion R in the waiting data set are randomly selected as training data set, and the other data are used as testing data set.
7. A modeling method for pedestrian crossing road decision based on SVM algorithm as claimed in claim 6 wherein the ratio R is set to 80%.
8. The modeling method for pedestrian crossing road decision based on SVM algorithm as claimed in claim 6, wherein during model test, the number of times of switching the viewing direction of the pedestrian at a certain test data point in the test data set and the duration of keeping the side view state are used as input variables, the predicted pedestrian decision type label is calculated and output by the SVM algorithm-based pedestrian crossing road decision model, if the predicted pedestrian decision type label is the same as the actual decision type label, the model test is successful for the test data point, otherwise, the model test fails.
9. The modeling method for pedestrian crossing road decision based on SVM algorithm of claim 8 wherein if the ratio of the number of successfully tested test data points to the total number of data points in the test data set exceeds a predetermined threshold P, the model accuracy is acceptable and the modeling is successful.
10. A modeling method for pedestrian crossing road decision based on SVM algorithm as claimed in claim 8 wherein said predetermined threshold P is set to 85%.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011570760.8A CN112766538A (en) | 2020-12-26 | 2020-12-26 | Decision modeling method for pedestrian crossing road based on SVM algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011570760.8A CN112766538A (en) | 2020-12-26 | 2020-12-26 | Decision modeling method for pedestrian crossing road based on SVM algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112766538A true CN112766538A (en) | 2021-05-07 |
Family
ID=75695954
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011570760.8A Pending CN112766538A (en) | 2020-12-26 | 2020-12-26 | Decision modeling method for pedestrian crossing road based on SVM algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112766538A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118015873A (en) * | 2024-01-09 | 2024-05-10 | 徐州瑞之然建设工程有限公司 | Intelligent traffic control system based on vehicle-road cooperation |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111178195A (en) * | 2019-12-18 | 2020-05-19 | 中国平安人寿保险股份有限公司 | Facial expression recognition method and device and computer readable storage medium |
| CN111985386A (en) * | 2020-08-17 | 2020-11-24 | 清华大学 | Method for identifying pedestrian illegal-passing behavior based on planned behavior theory |
-
2020
- 2020-12-26 CN CN202011570760.8A patent/CN112766538A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111178195A (en) * | 2019-12-18 | 2020-05-19 | 中国平安人寿保险股份有限公司 | Facial expression recognition method and device and computer readable storage medium |
| CN111985386A (en) * | 2020-08-17 | 2020-11-24 | 清华大学 | Method for identifying pedestrian illegal-passing behavior based on planned behavior theory |
Non-Patent Citations (2)
| Title |
|---|
| HONGJIA ZHANG: "Research on a Pedestrian Crossing Intention Recognition Model Based on Natural Observation Data", 《SENSORS》 * |
| 蔡逸超: "基于监控视频的行人头部姿态估计方法研究", 《中国优秀硕士论文电子期刊》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118015873A (en) * | 2024-01-09 | 2024-05-10 | 徐州瑞之然建设工程有限公司 | Intelligent traffic control system based on vehicle-road cooperation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7106664B2 (en) | Intelligent driving control method and device, electronic device, program and medium | |
| CN110248861B (en) | Guiding a vehicle using a machine learning model during vehicle maneuvers | |
| CN111919225B (en) | Training, testing, and validating autonomous machines using a simulated environment | |
| CN111369831A (en) | Road driving danger early warning method, device and equipment | |
| CN111582189B (en) | Traffic signal lamp identification method and device, vehicle-mounted control terminal and motor vehicle | |
| US20210389133A1 (en) | Systems and methods for deriving path-prior data using collected trajectories | |
| JP2023529959A (en) | Systems and methods for withdrawal prediction and triage assistance | |
| CN107909012B (en) | Real-time vehicle tracking detection method and device based on disparity map | |
| EP3920128A1 (en) | Operational design domains in autonomous driving | |
| US20230278587A1 (en) | Method and apparatus for detecting drivable area, mobile device and storage medium | |
| JP7454685B2 (en) | Detection of debris in vehicle travel paths | |
| CN117056153A (en) | Methods, systems, and computer program products for calibrating and verifying driver assistance systems and/or autopilot systems | |
| CN212570057U (en) | Road driving danger early warning device and equipment | |
| CN116266380A (en) | Environment data reconstruction method, device, system and storage medium | |
| US11209830B2 (en) | Safety aware automated governance of vehicles | |
| US20200189611A1 (en) | Autonomous driving using an adjustable autonomous driving pattern | |
| CN112766538A (en) | Decision modeling method for pedestrian crossing road based on SVM algorithm | |
| CN112735163B (en) | Method for determining static state of target object, road side equipment and cloud control platform | |
| WO2024098992A1 (en) | Vehicle reversing detection method and apparatus | |
| CN117612364A (en) | Method and device for detecting motor vehicle violation, electronic equipment and storage medium | |
| JP2023504506A (en) | perceptual error model | |
| CN105719313B (en) | A kind of motion target tracking method based on intelligent real-time video cloud | |
| CN111077893B (en) | Navigation method based on multiple vanishing points, electronic equipment and storage medium | |
| Tsu-Tian | Research on intelligent transportation systems in Taiwan | |
| CN118323198B (en) | Training and using method and device of decision model in automatic driving vehicle and vehicle |
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: 20210507 |
|
| RJ01 | Rejection of invention patent application after publication |