CN105868469A - Lane departure forewarning method based on perspective image and forewarning model construction method - Google Patents

Lane departure forewarning method based on perspective image and forewarning model construction method Download PDF

Info

Publication number
CN105868469A
CN105868469A CN201610184574.8A CN201610184574A CN105868469A CN 105868469 A CN105868469 A CN 105868469A CN 201610184574 A CN201610184574 A CN 201610184574A CN 105868469 A CN105868469 A CN 105868469A
Authority
CN
China
Prior art keywords
vehicle
lane
early
warning
under
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
Application number
CN201610184574.8A
Other languages
Chinese (zh)
Inventor
张洪丹
陈涛
李永利
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hunan University
Original Assignee
Hunan University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Hunan University filed Critical Hunan University
Priority to CN201610184574.8A priority Critical patent/CN105868469A/en
Publication of CN105868469A publication Critical patent/CN105868469A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00Computer-aided design [CAD]
    • G06F30/30Circuit design
    • G06F30/36Circuit design at the analogue level
    • G06F30/367Design verification, e.g. using simulation, simulation program with integrated circuit emphasis [SPICE], direct methods or relaxation methods
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00Three dimensional [3D] modelling, e.g. data description of 3D objects
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00Details relating to CAD techniques
    • G06F2111/10Numerical modelling

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Geometry (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Computer Graphics (AREA)
  • Microelectronics & Electronic Packaging (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Traffic Control Systems (AREA)
  • Image Processing (AREA)

Abstract

The invention relates to a lane departure forewarning method based on a perspective image and a forewarning model construction method. According to the forewarning method, whether a vehicle departs from a lane or not can be precisely judged on the basis of the included angle formed between lane lines and distance signal information between the vehicle gravity center and the longitudinal middle line of the lane in the perspective image. The forewarning method comprises the steps that firstly, the requirement of a system for the initial included angle formed between the left lane line and the right lane line is met by adjusting the installing angle of a camera; secondly, virtual simulation is conducted on a TLC model in an inverse perspective environment through Prescan/simulink software, and then the included angle formed between the left lane line and the right lane line and the distance that the vehicle gravity center departs from the lane center are simulated at different speeds and different yaw angles when the system sends out forewarning prompt, wherein the included angle and the distance are displayed in the perspective image; lastly, a three-dimensional forewarning model based on the lane line included angle and the distance when the vehicle departs from the lane at the different speeds and the different yaw angles is fitted. The forewarning model reminds a driver of the danger of departing from the lane 0.5 s ahead, that is, in-time forewarning is guaranteed, and meanwhile the phenomenon that frequent forewarning causes driving fatigue is avoided.

Description

A kind of lane departure warning method based on fluoroscopy images and Early-warning Model construction method
Technical field
The present invention relates to intelligent transportation field, particularly relate to a kind of deviation based on fluoroscopy images pre- Alarm method and Early-warning Model construction method.
Background technology
At present, automobile assistant driving technology is paid high attention to colleges and universities by Ge great automobile vendor, and car Deviation early warning system in road is driven when driver deviates original track due to fatigue driving as one The automobile assistant driving technology of the person's of sailing alarm, has attracted the concern of large quantities of automobile vendors, And it is the substantial amounts of manpower and materials of Innovation Input of this system.
Investigation statistics shows, on highway, the vehicle accident close to half deviates original track with vehicle Relevant, and most of vehicle deviates original track and is caused by fatigue driving.So developing a Can in the case of driver is unconscious will run-off-road time give driver's early warning remind system compel At the eyebrows and eyelashes, and the just right early warning moment is most important, is caused the most pre-by giving warning in advance Alert, driving fatigue, the generation that delayed early warning then can not well avoid traffic accident will be caused.
About the prediction policy of Lane Departure Warning System, mainly include the Early-warning Model of several classics, And relatively broad with the application of TLC model, but the application of TLC model is based on inverse fluoroscopy images, i.e. In image, lane line is parallel, and the fluoroscopy images that video camera is photographed, TLC model cannot obtain Well application.
Summary of the invention
For the problems referred to above, the present invention obtains TLC by the form of Prescan/simulink analog simulation A kind of new Early-warning Model that model is mapped in fluoroscopy images, thus conveniently carry out in fluoroscopy images Accurate early warning.It is inclined that the method for early warning of the present invention can carry out accurate track easily in fluoroscopy images From early warning, and save the complicated miscellaneous step of image conversion, save program runtime, improve Early warning efficiency.
Specifically, the present invention provides a kind of lane departure warning method based on fluoroscopy images, and it is special Levy and be, comprise the following steps that
Step 1, the model parameter of setting TLC lane departure warning model;
Step 2, carrying out emulation experiment modeling, modeling process includes: carry on simulating vehicle for Carry out the sensor of vehicle driving parameters measurement under inverse perspective mode, and under perspective mode, shoot car The video camera of forward image, regulation sensor measure field and the visual field of video camera, make the two each other Overlap, speed and traveling yaw angle, travel route and the width of road of vehicle is set;
Step 3, in simulated scenario, utilize sensor to obtain the driving parameters of vehicle under inverse perspective mode Carry out early warning judgement being supplied to TLC lane departure warning model, simulate in each friction speed, each In the case of individual different yaw angle, there is the situation of deviation in vehicle in travelling, and, utilize TLC track Vehicle deviation is reported to the police by deviation Early-warning Model based on the parameter under inverse perspective mode, and records early warning Sensor measures parameters under fluoroscopy images that moment shot by camera arrives and inverse perspective mode;
Left and right under early warning moment perspective mode in the case of step 4, acquisition friction speed, different yaw angle Lane line angle and cross track distance information;
Step 5, according to left and right lane line angle and vehicle's center of gravity and the track under deviation moment perspective mode The distance of center line and the velocity information of vehicle, set up three-dimensional Early-warning Model;
Step 6, according to set up three-dimensional Early-warning Model, by load video simulation verify its Early-warning Model Effectiveness;
Step 7, perspective mode image based on actual photographed, extract the lane line in perspective mode image Angle, and obtain the velocity information of vehicle, utilize described three-dimensional based on lane line angle and velocity information Early-warning Model carries out vehicle deviation early warning.
In a kind of preferred implementation, described step 2 includes:
Step 201: set up the driving path of vehicle in Prescan software, sets the width in track;
Step 202: carry the lane detection sensor-Lane required for Lane Departure Warning System Marker sensor and video camera, and regulate video camera and the position of sensor and angle so that camera The aspect ratio that in the picture photographed, sky is pre-set with system with the ratio of road keeps one Cause;
Step 203: setting speed and the width of vehicle, speed and width information according to vehicle solve car Yaw angle this moment.
In a kind of preferred implementation, described step 3 includes:
Step 301: be set under same speed, the yaw angle scope of simulating vehicle;
Step 302: simulating the vehicle early warning moment in the case of same speed, different yaw angle, record is pre- The fluoroscopy images in alert moment;
Step 303: the fluoroscopy images in the early warning moment remained is processed, and records same speed The left and right lane line angle in early warning moment and the cross track distance in units of pixel when degree, different yaw angle.
In a kind of preferred implementation, described step 1 includes: set by 0.5s before automotive run-off-road It is set to pre-warning time parameter T,Wherein, y0Represent vehicle's center of gravity away from wanting run-off-road line Lateral separation, w is the width of vehicle, and v is the absolute velocity of vehicle, and θ is the yaw angle of vehicle.
In a kind of preferred implementation, described step 203 includes: speed based on vehicle and width Information Pull simulink plug-in unit solves the yaw angle of vehicle.
In a kind of preferred implementation, described step 4 also includes by Vc6.0/Opencv lane line Extraction procedure extracts lane line angle and vehicle's center of gravity about the early warning moment based on the picture under perspective mode Range information away from track center line.
In a kind of preferred implementation, set yaw angle scope is θ ∈ [-10 °, 10 °], and vehicle Velocity interval be v ∈ [0,35m/s].
In a kind of preferred implementation, it is imitative that described method also includes utilizing Prescan software to be simulated Very, while calculating the position relationship of vehicle and lane line under the inverse perspective mode of image, output figure Vehicle under the perspective mode of picture and the position relationship of lane line.
On the other hand, the present invention provides a kind of structure lane departure warning based on fluoroscopy images model Method, it comprises the following steps that
Step 1, the model parameter of setting TLC lane departure warning model;
Step 2, carrying out emulation experiment modeling, modeling process includes: carry on simulating vehicle for Carry out the sensor of vehicle driving parameters measurement under inverse perspective mode, and under perspective mode, shoot car The video camera of forward image, regulation sensor measure field and the visual field of video camera, make the two each other Overlap, speed and traveling yaw angle, travel route and the width of road of vehicle is set;
Step 3, in simulated scenario, utilize sensor to obtain the driving parameters of vehicle under inverse perspective mode Carry out early warning judgement being supplied to TLC lane departure warning model, simulate in each friction speed, each In the case of individual different yaw angle, there is the situation of deviation in vehicle in travelling, and, utilize TLC track Vehicle deviation is reported to the police by deviation Early-warning Model based on the parameter under inverse perspective mode, and records early warning Sensor measures parameters under fluoroscopy images that moment shot by camera arrives and inverse perspective mode;
In the case of step 4, acquisition friction speed, different yaw angle, the early warning moment is against the left side under perspective mode Right lane wire clamp angle and cross track distance information, and by lane line angle under acquired inverse perspective mode and Cross track distance information is associated with the corresponding information in fluoroscopy images;
Step 5, according to left and right lane line angle and vehicle's center of gravity and the track under deviation moment perspective mode The distance of center line and the velocity information of vehicle, set up three-dimensional Early-warning Model;
Step 6, according to set up three-dimensional Early-warning Model, by load video simulation verify its Early-warning Model Effectiveness.
It should be noted that the cross track distance being previously mentioned in the present invention refers in vehicle's center of gravity and track Distance between heart line.
Beneficial effects of the present invention:
(1) present invention can carry out vehicle deviation early warning based on fluoroscopy images, and without by fluoroscopy images It is converted into inverse fluoroscopy images;
(2) threedimensional model set up can obtain good early warning effect under fluoroscopy images, and effect is equal to In classical TLC Early-warning Model;
(3) present invention reduces the amount of calculation of method for early warning, it is to avoid the loaded down with trivial details calculating to image Or conversion, save the sequential operation time;
Accompanying drawing explanation
Fig. 1 is inverse fluoroscopy images under normal circumstances and fluoroscopy images comparison diagram;
Fig. 2 show the embodiment of the present invention when carrying out early warning, between involved parameters Relation;
Fig. 3 shows embodiment of the present invention schematic diagram during vehicle yaw angle calculates;
Fig. 4 shows in fluoroscopy images, angle and the relation of range information;
Fig. 5 shows the three-dimensional Early-warning Model that the embodiment of the present invention provides.
Detailed description of the invention
Embodiment 1
Below with reference to accompanying drawing 1-5, the method for early warning provided in one embodiment of the invention is carried out in detail Describe in detail bright.
Method for early warning in the present invention is implemented under fluoroscopy images, and relevant lane departure warning system For the classical Early-warning Model of system is both for inverse fluoroscopy images.The new prediction policy of the present invention is Propose on the basis of TLC Early-warning Model, the TLC Early-warning Model under inverse fluoroscopy images " is translated " Become under fluoroscopy images based on left and right lane line angle and vehicle's center of gravity and track two kinds of information of center line distance Method for early warning, concrete grammar will will be described in detail below.
As it is shown in figure 1, the difference against fluoroscopy images with fluoroscopy images is, lane line in inverse perspective view Being parallel and distance (degree of depth) information can be extracted, and lane line is crossing in fluoroscopy images And range information is unknowable, and the picture that monocular-camera photographs is all fluoroscopy images, i.e. left and right car Diatom is certain angle, and vehicle is unknowable with the range information of surrounding objects, therefore, existing Method for early warning is mainly all based on what inverse fluoroscopy images was carried out, the invention mainly includes and sets up two kinds One " bridge " under picture state, by this seat " bridge ", the method for early warning in fluoroscopy images The effect that in inverse fluoroscopy images, TLC prediction policy reaches can be reached.
Present invention is generally directed to deviation system, it is provided that a kind of based on classical TLC method for early warning A kind of new method for early warning.Below, step the method is described in detail one by one.
Step 1, determine the TLC Early-warning Model of Lane Departure Warning System, take to deviate car at vehicle Before road, 0.5s gives and driver's early warning, i.e.Wherein T is the time parameter that gives warning in advance, y0Representing vehicle's center of gravity away from wanting the lateral separation cross track distance of run-off-road line, w is the width of vehicle Degree, v is the speed of vehicle, and θ is the yaw angle of vehicle.
It is illustrated in figure 2 between the parameters involved by TLC method for early warning under inverse fluoroscopy images Relation schematic diagram, the method for early warning under inverse fluoroscopy images relates generally to four factors, the width w of vehicle, The speed v of vehicle, yaw angle θ of vehicle, vehicle's center of gravity are away from lane line distance y0
For choosing the TLC method for early warning basis as the present invention of classics, main cause is that TLC is pre- It is commonplace that alarm method uses in Lane Departure Warning System, and can be examined by the corner of steering wheel Worry is entered, and improves the degree of accuracy of model, and can pre-set time T carry out early warning, reserved In the response time of driver, pilot model is also contemplated in the middle of method for early warning.
For the TLC method for early warning chosen--In pre-warning time T, used herein 0.5s。
Early-warning Model will give the necessity reminded with driver's early warning not before run-off-road line at vehicle Accommodating doubtful, driver so can be made to make operation in advance, it is to avoid vehicle deviates original traveling lane, make Become vehicle accident.But the value that the occurrence of threshold value T does not the most determine, if threshold value is excessive, the most not Being avoided that the generation of vehicle accident, threshold value is too small, then can frequently send early warning, be easily caused and drive The person's of sailing driving fatigue, data statistic analysis obtains when threshold value T=0.5s, can well take into account both.
Road surface and the ratio of sky in the picture that step 2, setting video camera photograph, and it is real to carry out emulation Testing modeling, carry out building of experiment scene by Prescan software, mainly on vehicle, carried experiment needs The sensor wanted, arranges the speed of vehicle and travels yaw angle, travel route and the width of road, By Prescan associating simulink plug-in unit, system early warning model algorithm is imported phantom afterwards;Should Step specifically includes:
Step 201: set up the yaw angle of the driving path of vehicle, i.e. vehicle in Prescan software, And set the width in track;
The reason of the modeling that selection carries out simulated environment in Prescan software is: in Prescan software It is provided with the sensor-Lane Marker sensor needed for Lane Departure Warning System, and can be convenient Build the scene required for system, include road, vehicle parameter, car speed, and the width in track Degree is set as 3.75m, and the range set of the yaw angle that vehicle travels is θ=[-10 °, 10 °], and every 1 Degree once emulates, and different yaw angles can cause the difference in early warning moment;
Step 202: carry the lane detection sensor-Lane required for Lane Departure Warning System Marker sensor and video camera, and regulate video camera and the position of sensor and angle so that camera The aspect ratio that in the picture photographed, sky is pre-set with system with the ratio of road keeps one Cause;
As it is shown on figure 3, be message processing module and message output module, i.e. Prescan in Prescan software When processing vehicle with lane line position relationship, process under inverse fluoroscopy images, and can lead to Cross and the mode of photographic head is installed on vehicle exports fluoroscopy images this moment.The advantage of Prescan software is just It is (vehicle to be included against the position relationship of (image) process vehicle under perspective mode with lane line Angle, range information with lane line) while, the car under perspective mode (image) can be exported With the position relationship of lane line.
Vehicle is installed Lane Marker sensor and Camera photographic head, and two biographies of regulation The angle of sensor and the length-width ratio of picture photographed, it is ensured that the visual field that two kinds of sensors photograph is identical.
Step 203: setting speed and the width of vehicle, speed and width information according to vehicle exist Simulink writes prediction policy algorithm and according to Lane Marker sensor transmissions by S function Information solves vehicle yaw angle this moment by writing S function;
During the associative simulation of Prescan Yu simulink, can be by the signal of sensor collection in Prescan Speed, vehicle yaw angle as the input signal in prediction policy, such as vehicle.In order to by these Signal is preferably combined with simulink module and simplifies simulink model, prediction policy module is led to Cross S function to write, and for the substantial amounts of information of Lane transmission, carry out information processing by S function, And finally calculate vehicle yaw angle this moment.It is concrete as it is shown on figure 3, pass through Lane Marker Intersection point between sensor scan line and longitudinal perpendicular bisector and the lane line of vehicle, calculates vehicle this moment Yaw angle.
Step 3, form by Prescan/simulink associative simulation, simulating vehicle is in same speed The automotive run-off-road early warning moment in the case of different yaw angles, and record the picture in this early warning moment, Lane line angle and car about the early warning moment is extracted afterwards by Vc6.0/Opencv lane line extraction procedure The range information of distance of centre of gravity track center line, and note down;This step specifically includes:
Step 301: be set under same speed, when such as speed is 5m/s, the driftage of simulating vehicle Angle range is θ ∈ [-10 °, 10 °], can meet the requirement at vehicle yaw angle in the range of this, and vehicle Velocity interval is v ∈ [0,35m/s], meets the rate request of highway;
Owing to deviation system is mainly applied on a highway, and the speed limit of highway is 120km/h, the speed arranging vehicle in phantom is v ∈ [0,35m/s], it is possible to meet highway Rate request, under same speed, it is ensured that before automotive run-off-road 0.5s give and driver's early warning On the premise of prompting, different yaw angles can cause the position of vehicle early warning persistently to change, it is contemplated that To the angle problem of yaw angle, in the range of yaw angle θ ∈ [-10 °, 10 °], substantially can meet vehicle to partially The requirement at boat angle, and during regulation vehicle left avertence, yaw angle is negative, during right avertence, vehicle yaw angle is just;
Step 302: simulate same speed by Prescan/simulink, different yaw angle situation is got off Early warning moment, when automotive run-off-road being detected in early warning program, by Pause function by program Stopping, and record early warning moment photographic plate, the yaw angle of vehicle is simulated once every once;
The early warning moment of automotive run-off-road is simulated, due to software emulation process only in Prescan software It is to show automotive run-off-road early warning by visual manner, is difficult to tell system accurately and starts early warning The first frame picture.Can be by adding Pause function in prediction policy S function, when program judges When going out automotive run-off-road line, termination program continues to run with, and records the fluoroscopy images in deviation moment, And under this specific speed, the yaw angle scope of simulating vehicle is θ ∈ [-10 °, 10 °], every 1 degree Simulation once, records 20 fluoroscopy images deviateing the moment under this speed altogether;
Step 303: by writing the saturating of the intact lane detection program early warning moment to remaining Visible image processes, and the track, left and right in deviation moment when recording under same speed different yaw angle Wire clamp angle and the vehicle's center of gravity distance (in units of pixel) between the center line of track.
In the case of same speed difference yaw angle, 20 run-off-road moment that simulation obtains Fluoroscopy images, by the most compiled lane detection program, detects each run-off-road moment saturating The left and right lane line angle theta of visible image and vehicle's center of gravity distance y between the longitudinal midline of track0, and right The two information makes a record, and angle is with range information as shown in Figure 4;
Step 4, by Prescan software change vehicle travel speed, car speed in the range of V ∈ [0,35m/s], emulates once every 5m/s, again records the left side in friction speed difference yaw angle moment Right lane wire clamp angle and vehicle's center of gravity range information between the center line of track:
Step 401: the travel speed that vehicle is different is set, be respectively provided with the speed of vehicle be 5m/s, 10m/s, 15m/s, 20m/s, 25m/s, 30m/s, 35m/s, and at each speed Imitating In the early warning moment in the case of emulation yaw angle θ ∈ [-10 °, 10 °], record early warning moment fluoroscopy images;
In factor influential for the early warning moment, speed and vehicle yaw angle be most important two because of Element, same yaw angle, when speed is big, the early warning moment shifts to an earlier date, speed hour, and the early warning moment is delayed, In order to meet the requirement at highway driving, the car speed of simulation is v ∈ [0,35m/s], every 5m/s Simulate once, the seriality of basic guarantee rate curve, and yaw angle is every 1 degree of simulation once.
Step 402: according to deviateing track wire clamp about the moment under lane detection Program extraction friction speed Range information between angle and vehicle's center of gravity and track center line.
The fluoroscopy images in the deviation moment for recording, by lane line extraction procedure, carries one by one Take lane line angle theta and vehicle's center of gravity distance y between the longitudinal midline of track about the deviation moment0, altogether Count 140 groups of data, record these data messages.
Step 5, distance according to left and right lane line angle and the vehicle's center of gravity in deviation moment with track center line Information, by MATLAB edit routine, sets up three-dimensional Early-warning Model;In this step.
Step 501: calculating vehicle is the track in run-off-road early warning moment under friction speed difference yaw angle Range information between wire clamp angle and vehicle and track center line.
140 groups of data messages are added up, and carries out differentiation, it is ensured that the corresponding vehicle that data are correct Speed, the corresponding 20 groups of deviation moment lane line angles of each speed and range information.
Step 502: write in MATLAB based on speed, left and right lane line angle, vehicle away from car The three-dimensional Early-warning Model drawing program of road three factors of center line distance, in order to preferably show 3-D graphic Effect, it is stipulated that car speed is just to left avertence, is negative to right avertence, and lane line angle is negative to left avertence, Just being to right avertence, vehicle's center of gravity is away from track longitudinal midline distance, and left avertence is just, right avertence is negative.Simulation Result as it is shown in figure 5, and identical in view of the equidirectional symbol of speed, do not have positive and negative point, institute With safety zone actual in Fig. 5 it is, after blue region is come with the face symmetry that speed is zero, More than blueness, red following part, for safety zone, other regions are prewarning area.
For the display that Lane Departure Warning System Early-warning Model based on fluoroscopy images is become apparent from, Using car speed, lane line angle, vehicle away from track longitudinal direction perpendicular bisector distance as the three of Early-warning Model Individual coordinate figure, and by the way of writing m file in MATLAB, by directly perceived for three-dimensional Early-warning Model Show.
Step 6, according to set up three-dimensional Early-warning Model, by load video simulation verify its Early-warning Model Effectiveness.
Step 7, once establish three-dimensional Early-warning Model, it is possible to pavement image based on actual photographed, Lane line angle in extraction image, and obtain the velocity information of vehicle, based on lane line angle and speed Three-dimensional Early-warning Model described in degree Information Pull carries out vehicle deviation early warning.
In the method for the present invention, in simulation process, by passing through sensor on simulating vehicle simultaneously Obtain the image under inverse perspective mode or the vehicle front image under result, and perspective mode, then Early warning is carried out based on the measurement result under inverse perspective mode by TLC Early-warning Model, and when recording early warning Carve the image under perspective mode, be then based on the image under the image zooming-out perspective mode under perspective mode Parameter (the lane line angle in image and cross track distance), then that the result under fluoroscopy images is saturating with inverse Result under visible image associates, it is possible to know, in the case of this speed and yaw angle, Under perspective mode, how to trigger early warning, after various speed and yaw angle are all simulated, Just can set up the three-dimensional Early-warning Model under fluoroscopy images, and then based on this three-dimensional Early-warning Model, permissible Directly carry out early warning judgement by fluoroscopy images.Saving calculating resource, response speed is fast, and efficiency is high.
Although the principle of the present invention having been carried out detailed retouching above in conjunction with the preferred embodiments of the present invention State, it should be appreciated by those skilled in the art that above-described embodiment is only the schematic realization to the present invention The explanation of mode, not comprises the restriction of scope to the present invention.It is right that details in embodiment is not intended that The restriction of the scope of the invention, without departing from the spirit and scope of the present invention, any based on this The equivalent transformation of inventive technique scheme, simple replacement etc. obviously change, and all fall within the present invention and protect Within the scope of protecting.

Claims (9)

1. a lane departure warning method based on fluoroscopy images, it is characterised in that include that step is such as Under:
Step 1, the model parameter of setting TLC lane departure warning model;
Step 2, carrying out emulation experiment modeling, modeling process includes: carry on simulating vehicle for Carry out the sensor of vehicle driving parameters measurement under inverse perspective mode, and under perspective mode, shoot car The video camera of forward image, regulation sensor measure field and the visual field of video camera, make the two each other Overlap, speed and traveling yaw angle, travel route and the width of road of vehicle is set;
Step 3, in simulated scenario, utilize sensor to obtain the driving parameters of vehicle under inverse perspective mode Carry out early warning judgement being supplied to TLC lane departure warning model, simulate in each friction speed, each In the case of individual different yaw angle, there is the situation of deviation in vehicle in travelling, and, utilize TLC track Vehicle deviation is reported to the police by deviation Early-warning Model based on the parameter under inverse perspective mode, and records early warning Sensor measures parameters under fluoroscopy images that moment shot by camera arrives and inverse perspective mode;
Left and right under early warning moment perspective mode in the case of step 4, acquisition friction speed, different yaw angle Lane line angle and cross track distance information;
Step 5, according to left and right lane line angle and vehicle's center of gravity and the track under deviation moment perspective mode The distance of center line and the velocity information of vehicle, set up three-dimensional Early-warning Model;
Step 6, according to set up three-dimensional Early-warning Model, by load video simulation verify its Early-warning Model Effectiveness;
Step 7, perspective mode image based on actual photographed, extract the lane line in perspective mode image Angle, and obtain the velocity information of vehicle, utilize described three-dimensional based on lane line angle and velocity information Early-warning Model carries out vehicle deviation early warning.
Lane departure warning method based on fluoroscopy images the most according to claim 1, its feature Being, described step 2 includes:
Step 201: set up the driving path of vehicle in Prescan software, sets the width in track;
Step 202: carry the lane detection sensor-Lane required for Lane Departure Warning System Marker sensor and video camera, and regulate video camera and the position of sensor and angle so that camera The aspect ratio that in the picture photographed, sky is pre-set with system with the ratio of road keeps one Cause;
Step 203: setting speed and the width of vehicle, speed and width information according to vehicle solve car Yaw angle this moment.
Lane departure warning method based on fluoroscopy images the most according to claim 1, its feature Being, described step 3 includes:
Step 301: be set under same speed, the yaw angle scope of simulating vehicle;
Step 302: simulating the vehicle early warning moment in the case of same speed, different yaw angle, record is pre- The fluoroscopy images in alert moment;
Step 303: the fluoroscopy images in the early warning moment remained is processed, and records same speed The left and right lane line angle in early warning moment and the cross track distance in units of pixel when degree, different yaw angle.
Lane departure warning method based on fluoroscopy images the most according to claim 1, its feature Being, described step 1 includes: 0.5s before automotive run-off-road is set as pre-warning time parameter T,Wherein, y0Representing vehicle's center of gravity away from wanting the lateral separation of run-off-road line, w is vehicle Width, v is the absolute velocity of vehicle, and θ is the yaw angle of vehicle.
Lane departure warning method based on fluoroscopy images the most according to claim 1, its feature Being, described step 203 includes: speed based on vehicle and width information utilize simulink plug-in unit Solve the yaw angle of vehicle.
Lane departure warning method based on fluoroscopy images the most according to claim 1, it is special Levying and be, described step 4 also includes by Vc6.0/Opencv lane line extraction procedure based on perspective mould Picture under formula extracts lane line angle and the vehicle's center of gravity distance letter away from track center line about the early warning moment Breath.
Lane departure warning method based on fluoroscopy images the most according to claim 1, its feature Being, set yaw angle scope is θ ∈ [-10 °, 10 °], and the velocity interval of vehicle is v∈[0,35m/s]。
Lane departure warning method based on fluoroscopy images the most according to claim 2, its feature Being, described method also includes utilizing Prescan software to be simulated emulation, at the inverse perspective mould of image Vehicle while calculating the position relationship of vehicle and lane line under formula, under the perspective mode of output image Position relationship with lane line.
9. the method building lane departure warning model based on fluoroscopy images, it includes that step is such as Under:
Step 1, the model parameter of setting TLC lane departure warning model;
Step 2, carrying out emulation experiment modeling, modeling process includes: carry on simulating vehicle for Carry out the sensor of vehicle driving parameters measurement under inverse perspective mode, and under perspective mode, shoot car The video camera of forward image, regulation sensor measure field and the visual field of video camera, make the two each other Overlap, speed and traveling yaw angle, travel route and the width of road of vehicle is set;
Step 3, in simulated scenario, utilize sensor to obtain the driving parameters of vehicle under inverse perspective mode Carry out early warning judgement being supplied to TLC lane departure warning model, simulate in each friction speed, each In the case of individual different yaw angle, there is the situation of deviation in vehicle in travelling, and, utilize TLC track Vehicle deviation is reported to the police by deviation Early-warning Model based on the parameter under inverse perspective mode, and records early warning Sensor measures parameters under fluoroscopy images that moment shot by camera arrives and inverse perspective mode;
In the case of step 4, acquisition friction speed, different yaw angle, the early warning moment is against the left side under perspective mode Right lane wire clamp angle and cross track distance information;
Step 5, according to left and right lane line angle and vehicle's center of gravity and the track under deviation moment perspective mode The distance of center line and the velocity information of vehicle, set up three-dimensional Early-warning Model;
Step 6, according to set up three-dimensional Early-warning Model, by load video simulation verify its Early-warning Model Effectiveness.
CN201610184574.8A 2016-03-28 2016-03-28 Lane departure forewarning method based on perspective image and forewarning model construction method Pending CN105868469A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610184574.8A CN105868469A (en) 2016-03-28 2016-03-28 Lane departure forewarning method based on perspective image and forewarning model construction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610184574.8A CN105868469A (en) 2016-03-28 2016-03-28 Lane departure forewarning method based on perspective image and forewarning model construction method

Publications (1)

Publication Number Publication Date
CN105868469A true CN105868469A (en) 2016-08-17

Family

ID=56625159

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610184574.8A Pending CN105868469A (en) 2016-03-28 2016-03-28 Lane departure forewarning method based on perspective image and forewarning model construction method

Country Status (1)

Country Link
CN (1) CN105868469A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106774346A (en) * 2017-02-10 2017-05-31 苏州瑞得恩光能科技有限公司 For the positioner and its localization method of solar panel sweeping robot
CN107598927A (en) * 2017-10-20 2018-01-19 苏州瑞得恩光能科技有限公司 Solar panel sweeping robot straight trip decision maker and its decision method
CN110006665A (en) * 2019-04-22 2019-07-12 中国汽车工程研究院股份有限公司 A kind of indoor lane departure warning test method using virtual lane line
CN110196062A (en) * 2019-06-27 2019-09-03 成都圭目机器人有限公司 A kind of air navigation aid of one camera tracking lane line

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6725168B2 (en) * 2000-06-14 2004-04-20 Sumitomo Rubber Industries, Ltd. Vehicle/tire performance simulating method
CN103738243A (en) * 2013-10-29 2014-04-23 惠州华阳通用电子有限公司 Early warning method for lane departure
CN105620477A (en) * 2014-10-28 2016-06-01 奇瑞汽车股份有限公司 Lane departure early warning control method for vehicle
CN107194342A (en) * 2017-05-16 2017-09-22 西北工业大学 Method for detecting lane lines based on inverse perspective mapping

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6725168B2 (en) * 2000-06-14 2004-04-20 Sumitomo Rubber Industries, Ltd. Vehicle/tire performance simulating method
CN103738243A (en) * 2013-10-29 2014-04-23 惠州华阳通用电子有限公司 Early warning method for lane departure
CN105620477A (en) * 2014-10-28 2016-06-01 奇瑞汽车股份有限公司 Lane departure early warning control method for vehicle
CN107194342A (en) * 2017-05-16 2017-09-22 西北工业大学 Method for detecting lane lines based on inverse perspective mapping

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
张洪丹: ""车道偏离预警系统车道线检测及预警策略算法研究"", 《万方学位论文》 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106774346A (en) * 2017-02-10 2017-05-31 苏州瑞得恩光能科技有限公司 For the positioner and its localization method of solar panel sweeping robot
CN106774346B (en) * 2017-02-10 2023-11-24 苏州瑞得恩光能科技有限公司 Positioning device and positioning method for solar panel cleaning robot
CN107598927A (en) * 2017-10-20 2018-01-19 苏州瑞得恩光能科技有限公司 Solar panel sweeping robot straight trip decision maker and its decision method
CN110006665A (en) * 2019-04-22 2019-07-12 中国汽车工程研究院股份有限公司 A kind of indoor lane departure warning test method using virtual lane line
CN110196062A (en) * 2019-06-27 2019-09-03 成都圭目机器人有限公司 A kind of air navigation aid of one camera tracking lane line
CN110196062B (en) * 2019-06-27 2022-03-25 成都圭目机器人有限公司 Navigation method for tracking lane line by single camera

Similar Documents

Publication Publication Date Title
CN112198859B (en) Method, system and device for testing automatic driving vehicle in vehicle ring under mixed scene
CN109884916A (en) A kind of automatic Pilot Simulation Evaluation method and device
CN105045649B (en) A kind of intelligent automobile emulation mode and device
CN112997060A (en) Method and system for modifying a control unit of an autonomous vehicle
CN112224211A (en) Driving simulation system based on multi-autonomous-body traffic flow
CN105868469A (en) Lane departure forewarning method based on perspective image and forewarning model construction method
CN106599767A (en) Lane boundary detection data generation in virtual environment
Nentwig et al. Hardware-in-the-loop testing of computer vision based driver assistance systems
CN110569602B (en) Data acquisition method and system for unmanned vehicle
CN106064587A (en) Multi-lane vehicle distance recognition method and device based on license plate recognition
WO2022110737A1 (en) Vehicle anticollision early-warning method and apparatus, vehicle-mounted terminal device, and storage medium
CN104574993B (en) A kind of method of road monitoring and device
Olaverri-Monreal et al. Collaborative approach for a safe driving distance using stereoscopic image processing
WO2024016877A1 (en) Roadside sensing simulation system for vehicle-road collaboration
CN109917359A (en) Robust vehicle distances estimation method based on vehicle-mounted monocular vision
CN113918615A (en) Simulation-based driving experience data mining model construction method and system
CN107481291A (en) Traffic monitoring model calibration method and system based on mark dotted line physical coordinates
CN111160132B (en) Method and device for determining lane where obstacle is located, electronic equipment and storage medium
CN106228531A (en) Automatic vanishing point scaling method based on horizon search and system
EP4302165A1 (en) Instantiating objects in a simulated environment based on log data
CN110727269B (en) Vehicle control method and related product
CN116105712A (en) Road map generation method, reinjection method, computer device and medium
Yan Vehicle safety-assisted driving technology based on computer artificial intelligence environment
CN114820504A (en) Method and device for detecting image fusion deviation, electronic equipment and storage medium
Zhang et al. Research on unmanned system environment perception system methodology

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20160817