CN109949594A - Real-time traffic light recognition method - Google Patents
Real-time traffic light recognition method Download PDFInfo
- Publication number
- CN109949594A CN109949594A CN201910354808.2A CN201910354808A CN109949594A CN 109949594 A CN109949594 A CN 109949594A CN 201910354808 A CN201910354808 A CN 201910354808A CN 109949594 A CN109949594 A CN 109949594A
- Authority
- CN
- China
- Prior art keywords
- traffic lights
- information
- image
- lane
- road
- 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
Abstract
The present invention provides a kind of real-time traffic lamp recognition methods, comprising: obtains traffic lights information, and itself and original map are carried out fusion treatment, generate mark map；The Lane id in section Road id corresponding before stop line, lane is associated with traffic lights information respectively, generates the first and/or second related information table；According to the current location of vehicle and mark map, the Road id in the section and/or Lane id in lane is determined；According to the Road id in section and the first related information table and/or the Lane id and the second related information table in lane, when from vehicle apart from stop line within a preset range when, section before judging stop line whether there is traffic lights information, and in the presence of, obtain absolute location information of the lamp plate frame in mark map, and the picture element position information being converted into the first image；According to deep learning algorithm, detection and classification processing are carried out to the first ROI region, obtain the state of the corresponding traffic lights of id of traffic lights.The accuracy and speed of visual identity is improved as a result,.
Description
Technical field
The present invention relates to technical field of data processing more particularly to a kind of real-time traffic light recognition methods.
Background technique
Intelligent driving technology is hot topic in recent years, is alleviating traffic congestion, is improving road safety, reduction air
The fields such as pollution, can bring subversive change.Traffic lights identification is the essential important component of intelligent driving system,
The correct identification of traffic light signals plays key effect in outdoor safety navigation to intelligent driving system.Therefore, intelligent driving
System how quickly and accurately to identify traffic lights position, color and how reasonably decision start and stop become researcher and paying close attention to
Direction.
It can be divided mainly into the convenient method based on communications protocol using the recognition methods of more traffic lights at present and be based on
Map and two kinds of the traffic light recognition method of positioning.
Convenient method based on communications protocol is simply commonly exactly V2X, refers to and builds between traffic lights and automatic Pilot vehicle
Vertical communication.Fill an emission source on traffic lights first, emission source can externally emit the status information of traffic lights, recipient according to
The communication interface appointed in advance receives in real time emits the signal of coming at traffic lights, last automatic Pilot unmanned vehicle receives color
After signal, according to actual road conditions start and stop.The communication distance of V2X is up to a km, can be almost without any intrusively by traffic
Lamp information accurately passes to vehicle.But a large amount of hardware supported is needed, and cause cost too high, it can not be widely
It applies in actual environment scene.
Traffic light recognition method based on map and positioning improves traffic by means of accurate 3D map and self align technology
The reliability of lamp state recognition mainly includes high-precision map label, the region (region of interest, ROI) interested life
At with visual traffic lamp inspection survey and classify three parts.Firstly, traffic lights absolute position is labeled in map, then by by ground
The absolute position marked in figure is transformed into image-region, extracts only relevant to the traffic lights in vehicle-mounted camera captured image
ROI region, finally, carrying out the detection and classification of visual traffic lamp in ROI region.But the traffic based on map and positioning
The difficult point of lamp recognition methods is real-time and robustness, and existing such algorithm can not fully meet platform real-time and precision
Robustness.Secondly, reasonable map agreement, real-time map inquiry mode and start and stop decision is not known in existing algorithm, and single take the photograph
As head is in the limitation of range or length, so that such method is unable to satisfy in driving conditions to the detection visual field of traffic lights and model
It encloses, the problem of crossing the border is frequently present of.
Summary of the invention
The purpose of the embodiment of the present invention is that a kind of real-time traffic light recognition method is provided, to solve friendship in the prior art
Robustness when logical lamp identification is not high and problem of crossing the border.
To solve the above problems, in a first aspect, the present invention provides a kind of real-time traffic light recognition method, the method
Include:
Obtain traffic lights information；The traffic lights information includes the absolute of the lamp plate frame for identifying id and traffic lights of traffic lights
Location information；
The traffic lights information and original map are subjected to fusion treatment, generate mark map；The mark map includes
A plurality of road；Every road includes multiple road section informations；Each road section information includes the Road id in section and multiple
Lane information；Each lane information includes the Lane id in lane；
Section Road id corresponding before stop line and the traffic lights information are associated, the first related information is generated
Table, and/or, the Lane id in corresponding lane before the stop line and the traffic lights information are associated, second is generated and closes
Join information table；
Under steam, current location and the first image of vehicle are obtained；The first image is adopted including the use of the first image
The image of the traffic lights of acquisition means acquisition；
According to the current location of the vehicle and the mark map, the Road in the corresponding section in the current location is determined
The id and/or Lane id in lane；
It is corresponding according to the current location when at a distance from the current location with stop line in preset detection range
Section Road id and the first related information table, and/or, according to the lane in the corresponding section in the current location
Lane id and the second related information table, the section before judging stop line whether there is traffic lights information；
When there are traffic lights information, absolute location information of the lamp plate frame in the mark map is obtained；
The absolute location information is converted into picture element position information of the lamp plate frame in the first image；
According to the image for the traffic lights that the picture element position information and the first image acquisition device acquire, first is determined
ROI region；
According to deep learning algorithm, detection and classification processing are carried out to first ROI region, obtain the traffic lights
The state of the corresponding traffic lights of id.
In one possible implementation, the location information of the lamp plate frame of the traffic lights includes the lamp plate frame of traffic lights
Height of the longitude and latitude and each angle point of four angle points apart from ground.
In one possible implementation, the first image acquisition device is focal length camera.
In one possible implementation, the method also includes:
In current location and the first image for obtaining vehicle, the second image is obtained；Second image is including the use of
The image of the traffic lights of two image acquisition devices.
In one possible implementation, the method also includes:
When picture element position information of the lamp plate frame in the first image crosses the border；
The absolute location information of the lamp plate frame is converted into location of pixels of the lamp plate frame in second image
Information；
It is adopted according to picture element position information of the lamp plate frame in second image and second image collecting device
The image of the traffic lights of collection determines the second ROI region；
According to deep learning algorithm, detection and classification processing are carried out to second ROI region, obtain the traffic lights
The state of the corresponding traffic lights of id.
In one possible implementation, second image collector is set to short focus camera.
In one possible implementation, after the method further include:
Obtain vehicle current velocity information and acceleration information；
According to the current location of vehicle and the mark map, calculating vehicle is currently at a distance from stop line；
According to the velocity information, the acceleration information, the distance and the traffic lights state, calculate vehicle exist
The state at crossing.
Second aspect, the present invention provides a kind of equipment, including memory and processor, the memory is for storing journey
Sequence, the processor are used to execute any method of first aspect.
The third aspect, the present invention provides a kind of computer program products comprising instruction, when the computer program produces
When product are run on computers, so that the computer executes the method as described in first aspect is any.
Fourth aspect, the present invention provides a kind of computer readable storage medium, on the computer readable storage medium
It is stored with computer program, the method as described in first aspect is any is realized when the computer program is executed by processor.
By applying real-time traffic light recognition method provided by the invention, it is only necessary to two high-definition camera sensors,
It is at low cost；Sensor installation is simple, and the combination of long short focus fully meets the detection range and distance of traffic lights；Map inquiry side
Formula can obtain traffic light signals at an arbitrary position；The combination of high definition map label and vision-based detection range had both improved vision knowledge
Other precision, also improves speed, is able to satisfy the unmanned logistic car in market, unmanned road sweeper and unmanned passenger car completely to traffic lights
The real-time demand of identification；Reasonable start and stop decision not only alleviates the loss of unmanned vehicle, also gives the experience of occupant comfort.
Detailed description of the invention
Fig. 1 is the real-time traffic light recognition method flow diagram that the embodiment of the present invention one provides；
Fig. 2 is the absolute location information schematic diagram for the lamp plate frame that the embodiment of the present invention one provides；
Fig. 3 is the schematic diagram for the road section information that the embodiment of the present invention one provides；
Fig. 4 A is the corresponding first ROI region schematic diagram of Road id that the embodiment of the present invention one provides；
Fig. 4 B is the corresponding first ROI region schematic diagram of Lane id that the embodiment of the present invention one provides；
Fig. 5 is the change lamp stagnation of movement strategy schematic diagram that the embodiment of the present invention one provides.
Specific embodiment
The application is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched
The specific embodiment stated is used only for explaining related invention, rather than the restriction to the invention.It also should be noted that for just
Part relevant to related invention is illustrated only in description, attached drawing.
It should be noted that in the absence of conflict, the features in the embodiments and the embodiments of the present application can phase
Mutually combination.The application is described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.
Fig. 1 is the real-time traffic light recognition method flow diagram that the embodiment of the present invention one provides.This method is applied
In vehicle, it is particularly possible to apply in automatic driving vehicle.As shown in Figure 1, method includes the following steps:
Step 101, traffic lights information is obtained；Traffic lights information includes the lamp plate frame of the mark id and traffic lights of traffic lights
Absolute location information.
Specifically, the absolute location information of the lamp plate frame of traffic lights include four angle points of lamp plate frame of traffic lights longitude,
The height of latitude and each angle point apart from ground.
In map acquisition phase, can use map acquisition equipment and carry out map acquisition, example and it is non-limiting, which adopts
Collection equipment can be automatic driving vehicle itself, be also possible to intelligent robot, and can also be other has the function of map acquisition,
For example sweeper of the logistic car of automatic Pilot, automatic cleaning etc. is other when executing one's work, can carry out map by the way
The automatic driving vehicle of acquisition.
Example and limit, can use map acquisition equipment on global positioning system (Global Positioning
System, GPS), Global Satellite Navigation System (Global Navigation Satellite System, GNSS) etc. handed over
The acquisition of the location information of the lamp plate frame of logical lamp.The absolute location information of collected traffic lights lamp plate frame is referring to fig. 2.
Step 102, traffic lights information and original map are subjected to fusion treatment, generate mark map；Marking map includes
A plurality of road；Every road includes multiple road section informations；Each road section information includes the Road id and multiple lanes letter in section
Breath；Each lane information includes the Lane id in lane.
Specifically, traffic lights information can be marked with agreement into original map, for example, can be ordered with the id of traffic lights
Name, by each angle point of the lamp plate frame of traffic lights and longitude, latitude and the elevation information at angle point center with the shape of coordinate x, y, z
Formula is marked into original map.
Step 103, the Road id in section corresponding before stop line and traffic lights information are associated, generate first and closes
Join information table, and/or, lane Lane id corresponding before stop line and traffic lights information are associated, the second association letter is generated
Cease table.
Specifically, may include multiple sections (road) information, each road section information includes section on a long road
Road id and multiple lanes (lane) information, each lane information include the Lane id in lane.In multiple road section informations,
Before stop line (stopline) on corresponding section, the Road id in traffic lights information and section is associated, available one
A related information table, referred to as the first related information table.
Wherein, Road id can be associated with single or multiple traffic lights, depend primarily on the property of traffic intersection, it is assumed that
Traffic intersection is crossroad, and for vehicle when by crossroad, there are three lamps, then on the Road id where vehicle there are three associations
Lamp, example and it is non-limiting, the first related information table may refer to table 1.
Table 1
Assuming that vehicle at the parting of the ways, there are three lanes, then lane Lane id is associated with traffic lights information, example
And it is non-limiting, the second related information table is referring to table 2:
Table 2
Equipment is acquired by map as a result, obtains related information table, when map acquires the non-automatic driving vehicle of equipment itself
When, the mark map and related information table can be got from server, also be can receive other map acquisition equipment and are sent
Mark map and related information table.
Step 104, under steam, current location and the first image of vehicle are obtained；First image is including the use of the first figure
As the image for the traffic lights that acquisition device acquires.
In one embodiment, when vehicle carry out when driving, can by GPS obtain current location, by itself
First image collecting device obtains multiple image.
Subsequent, vehicle can use the processing unit of itself, carries out the processing such as noise reduction, filtering to each frame image, obtains
Treated image, according to the property of traffic intersection, when there are one or more traffic lights in a certain frame treated image
When image, which is known as the first image.
In another embodiment, vehicle in the process of moving, while can use itself the second image collecting device
Getting frame image is handled based on identical noise reduction, filtering etc., the image that obtains that treated, according to the property of traffic intersection, when certain
When there is the image of one or more traffic lights in one frame treated image, which is known as the second image.
It is subsequent, the selection of traffic scene can be carried out by the first image and the second image.
Wherein, the first image collector is set to focal length camera, and the second image collector is set to short focus camera.
Step 105, according to the current location of vehicle and mark map, the Road id in the corresponding section in current location is determined
And/or the Lane id in lane.
Wherein, the current location of vehicle includes longitude and latitude of the vehicle in world coordinate system, the position can with coordinate,
For example the form of (longitude, latitude) embodies.
Specifically, a road is made of multiple sections, and each section is one in the road in mark map
Section, vehicle include the image information of traffic lights in the process of moving, in acquired image.
It according to the position of vehicle, is inquired in mark map, determines the corresponding section in the position of vehicle, and determining should
The Road id in section, alternatively, the lane Lane id in the section is determined, alternatively, determining the Road id and lane Lane in the section
id。
Step 106, corresponding according to current location when at a distance from current location with stop line in preset detection range
Section Road id and the first related information table, and/or, according to the Lane id in the lane in the corresponding section in current location and
Second related information table, the section before judging stop line whether there is traffic lights information.
Specifically, obtaining traffic light signals within the specified range, for example obtain in 100 meters in vehicle travel process
Traffic light signals.
In one embodiment, when final stage road section length is less than 100 meters, vehicle can only obtain within the scope of 50 meters
Traffic light signals, it is therefore possible to use searching method in advance across section (road), i.e., according to the corresponding section of current vehicle position
Road id searches the Road id in the section in the front of current vehicle position whether in the first related information table, if vehicle
Not in the first related information table, vehicle continues to move forward the Road id in the front of current location, if it is present executing
Step 107.
Alternatively, the front of current vehicle position can also be searched with the Lane id in the corresponding section in enquiring vehicle current location
Section Lane id whether in the second related information table, if the Lane id in the section in the front of current vehicle position is not
In the second related information table, vehicle continues to move forward, if it is present executing step 107.
In another embodiment, when vehicle receives traffic light signals in a flash before stop line, due to inertia,
Stop line may have been gone out, and distance just stops a little, and vehicle is in crossroad at this time, but on the section of crossroad simultaneously
Traffic lights information is not identified, in order to solve this problem, can detect certain distance forward, i.e., according to the current location pair of vehicle
The Road id in the section answered inquires the Road id in a upper section for current road segment whether in the first related information table, such as
Fruit exists, and thens follow the steps 107.
Alternatively, inquiring a upper section for current road segment according to the Lane id in the corresponding section in current location of vehicle
Lane id whether in the second related information table, if it is present execute step 107.
The next of current road segment is inquired if the Road id in the corresponding section of current vehicle position is 88 referring to Fig. 3
The Road id in a section, as can be seen from the figure: next section of Road id 88 is Road id 89.Then judge Road
Id 89 whether there is traffic lights information, from the figure 3, it may be seen that Road id 89 corresponds to stop line, there are traffic lights information, therefore, when
It is available to arrive traffic lights information when vehicle is in Road id 88.
If the Road id in the corresponding section of current vehicle position is Road id 111, the first related information table is inquired,
Road id 111 is not in related information table.At this point it is possible to the Road id in a upper section of Road id 111 is inquired, and
Judge that the Road id 89 in a upper section whether there is traffic lights information, from the figure 3, it may be seen that Road id 89 corresponds to stop line,
There are traffic lights information, therefore, also available to arrive traffic lights information when vehicle is in Road id 111.
Correspondingly, being made whether that there are the judgements of traffic lights information about by Lane id and the second related information table
When, it is made whether to above by Road id and the first related information table there are the judgement of traffic lights information is similar, herein no longer
It repeats.
Step 107, when there are traffic lights information, absolute location information of the lamp plate frame in mark map is obtained.
Specifically, can obtain traffic lights from the first or second related information table when there are traffic lights information and exist
Mark the absolute location information in map.
Wherein, when being judged by Road id and the first related information table, the quantity of traffic lights at this time be can be
One or more, A referring to fig. 4.When being judged by Lane id and the second related information table, traffic lights at this time
Quantity is one, referring to fig. 4 B.When being judged by the two combination, two kinds of knots in Fig. 4 A and Fig. 4 B can be obtained simultaneously
Fruit.
Step 108, absolute location information is converted into picture element position information of the lamp plate frame in the first image.
Specifically, the internal reference (optical center, focal length, principal point etc.) and outer ginseng (spin matrix peace of focal length camera can be passed through
The amount of shifting to etc.) coordinate conversion is carried out, absolute location information of the lamp plate frame in mark map is converted to it in image coordinate system
Position, that is, the picture element position information being converted into the first image.Specific conversion method is the prior art, herein not
It repeats again.
Step 109, according to the image of picture element position information and the traffic lights of the first image acquisition device, first is determined
ROI region.
Since there are measurement error, there is also errors in coordinate transform process in the mark map of generation, therefore, lamp plate frame
The two-dimentional frame of the projection generated in the first image cannot be directly as the position frame of traffic lights, and the projection two dimension frame is in the first image
It is upper to be located near traffic lights physical location.Larger rectangle can be chosen as the first area ROI centered on the projection two dimension frame
Domain.
Specifically, the application uses focal length camera and short focus camera, and when carrying out ROI region selection, preferential selection length
Burnt camera, A and 4B still when projection is crossed the border, is switched to short focus camera referring to fig. 4.
Wherein, projection, which is crossed the border, refers to projection frame not in image range.For example, to the position of the lamp plate frame in traffic lights information
After information carries out a series of coordinate conversions, the coordinate in the first image is obtained, it is assumed that the first image resolution is 1280x720, i.e.,
The range of x is 0-1279, and the range of y is 0-719, if lamp plate frame is converted by coordinate, obtained coordinate is not in the first image
In the range of, referred to as projection is crossed the border.Alternatively, defining boundary sizes is 100 pixels, i.e. the range of x becomes 100-1179, y's
Range is 100-619, if lamp plate frame is converted by coordinate, obtained coordinate is not within the scope of this in the first image, also referred to as
It crosses the border for projection.The boundary sizes can be defined according to self-demand.
In vehicle travel process, when by the first related information table, the quantity of obtained traffic lights be it is multiple, such as 3
When a, the switchover policy of focal length camera and short focus camera can be set are as follows: when there is a throwing in the projection frame of 3 traffic lights
When shadow crosses the border, short focus camera is switched to；Alternatively, the projection frame of 3 traffic lights, which all projects to have crossed the border, is just switched to short focus
Camera.Specific switchover policy can be set according to actual needs, and the application does not limit this.
Step 110, according to deep learning algorithm, detection and classification processing is carried out to the first ROI region, obtain traffic lights
The state of the corresponding traffic lights of id.
Specifically, the training of deep learning network can be carried out in advance, prepare a large amount of training picture using prior, those
Traffic lights and color state in picture are all marked, using picture and the corresponding information marked as the defeated of deep learning network
Enter, recursive training is carried out to the weighting parameter in network, the model trained can be directly in the picture not being labeled
Traffic lights is detected and is classified.
Wherein, traffic light status is there are four types of type: red, yellow, and green and black.
Similarly, when picture element position information of the lamp plate frame in the first image crosses the border；
The absolute location information of lamp plate frame is converted into picture element position information of the lamp plate frame in the second image；
According to the traffic lights of picture element position information and second image acquisition device of the lamp plate frame in the second image
Image determines the second ROI region；
According to deep learning algorithm, detection is carried out to the second ROI region and classification processing, the id for obtaining traffic lights are corresponding
The state of traffic lights.
Further, after step 110, further includes:
Obtain vehicle current velocity information and acceleration information；
According to the current location of vehicle and mark map, calculating vehicle is currently at a distance from stop line；
According to the state of velocity information, acceleration information, distance and traffic lights, vehicle is calculated in the state at crossing.
Specifically, can get the velocity information of vehicle by wheel speed meter when for automatic driving vehicle, can pass through
Acceleration transducer gets the acceleration information of vehicle, can be calculated according to the current location of vehicle and the position of stop line
Vehicle is at a distance from stop line.
It is subsequent, it can the state according to the traffic lights received and the distance decision down time from vehicle distance stopline
And position, at this point it is possible to solve two problems: the quick high accuracy of the detection of 1. traffic lights and classification provides for real-time high-efficiency decision
It ensures, automatic driving vehicle is to traffic lights color change quick response, and when traffic lights greening starts at once, with manual operation phase
Than without extra delay.2. may stop the problem to after stop line for vehicle, decision also takes corresponding measure, this measure
Also the comfort level (will not bring to a halt) stopped when meeting red light from vehicle has been ensured.Intelligent driving bus location real-time monitoring is from vehicle distance
Whether the distance of lane line, travel speed, acceleration calculate to cross in the 3s of amber light and stop when encountering green light just flavescence lamp
Otherwise only line, slows down from vehicle if it can, meeting amber light from vehicle directly crosses stop line and prepares to stop before stop line.Referring to figure
5, when apart from stop line 15m, traffic lights becomes amber light, this brief acceleration is 0.5m/s, travel speed 50km/h, leads to
Calculating is crossed, needs 2.22s that can run over stop line, is i.e. amber light can cross safely before becoming red light, then not parking.The decision
It can not only prevent from crossing stop line from vehicle emergency stop, it is also ensured that very comfortable parking impression.
The real-time traffic light recognition method provided by the application embodiment of the present invention one, it is only necessary to two high-definition cameras
Sensor, it is at low cost；Sensor installation is simple, and the combination of long short focus fully meets the detection range and distance of traffic lights；Ground
Figure inquiry mode can obtain traffic light signals at an arbitrary position；The combination of high definition map label and vision-based detection range, was both improved
The precision of visual identity, also improves speed, is able to satisfy the unmanned logistic car in market, unmanned road sweeper and unmanned passenger car completely
To the real-time demand of traffic lights identification；Reasonable start and stop decision not only alleviates the loss of unmanned vehicle, also to occupant comfort
Experience.
Second embodiment of the present invention provides a kind of equipment, including memory and processor, memory is deposited for storing program
Reservoir can be connect by bus with processor.Memory can be nonvolatile storage, such as hard disk drive and flash memory, storage
Software program and device driver are stored in device.Software program is able to carry out the above method provided in an embodiment of the present invention
Various functions；Device driver can be network and interface drive program.Processor is for executing software program, the software journey
Sequence is performed, and can be realized method provided in an embodiment of the present invention.
The embodiment of the present invention three provides a kind of computer program product comprising instruction, when computer program product is being counted
When being run on calculation machine, so that computer executes the method that the embodiment of the present invention one provides.
The embodiment of the present invention four provides a kind of computer readable storage medium, is stored on computer readable storage medium
Computer program realizes the method that the embodiment of the present invention one provides when computer program is executed by processor.
Professional should further appreciate that, described in conjunction with the examples disclosed in the embodiments of the present disclosure
Unit and algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, hard in order to clearly demonstrate
The interchangeability of part and software generally describes each exemplary composition and step according to function in the above description.
These functions are implemented in hardware or software actually, the specific application and design constraint depending on technical solution.
Professional technician can use different methods to achieve the described function each specific application, but this realization
It should not be considered as beyond the scope of the present invention.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can be executed with hardware, processor
The combination of software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only memory
(ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technical field
In any other form of storage medium well known to interior.
Above specific embodiment has carried out further in detail the purpose of the present invention, technical scheme and beneficial effects
Illustrate, it should be understood that the above is only a specific embodiment of the invention, the protection model that is not intended to limit the present invention
It encloses, all within the spirits and principles of the present invention, any modification, equivalent substitution, improvement and etc. done should be included in the present invention
Protection scope within.
Claims (10)
1. a kind of real-time traffic light recognition method, which is characterized in that the described method includes:
Obtain traffic lights information；The traffic lights information includes the absolute position of the mark id of traffic lights and the lamp plate frame of traffic lights
Information；
The traffic lights information and original map are subjected to fusion treatment, generate mark map；The mark map includes a plurality of
Road；Every road includes multiple road section informations；Each road section information include section Road id and multiple lanes
Information；Each lane information includes the Lane id in lane；
Section Road id corresponding before stop line and the traffic lights information are associated, the first related information table is generated,
And/or be associated the Lane id in corresponding lane before the stop line and the traffic lights information, generate the second association
Information table；
Under steam, current location and the first image of vehicle are obtained；The first image is including the use of the first image collector
Set the image of the traffic lights of acquisition；
According to the current location of the vehicle and the mark map, the Road id in the corresponding section in the current location is determined
And/or the Lane id in lane；
When at a distance from the current location with stop line in preset detection range, according to the corresponding road in the current location
The Road id and the first related information table of section, and/or, according to the Lane in the lane in the corresponding section in the current location
Id and the second related information table, the section before judging stop line whether there is traffic lights information；
When there are traffic lights information, absolute location information of the lamp plate frame in the mark map is obtained；
The absolute location information is converted into picture element position information of the lamp plate frame in the first image；
According to the image for the traffic lights that the picture element position information and the first image acquisition device acquire, the first ROI is determined
Region；
According to deep learning algorithm, detection and classification processing are carried out to first ROI region, obtain id pairs of the traffic lights
The state for the traffic lights answered.
2. the method according to claim 1, wherein the location information of the lamp plate frame of the traffic lights includes traffic
Height of the longitude and latitude and each angle point of four angle points of lamp plate frame of lamp apart from ground.
3. the method according to claim 1, wherein the first image acquisition device is focal length camera.
4. the method according to claim 1, wherein the method also includes:
In current location and the first image for obtaining vehicle, the second image is obtained；Second image is including the use of the second figure
As the image for the traffic lights that acquisition device acquires.
5. according to the method described in claim 4, it is characterized in that, the method also includes:
When picture element position information of the lamp plate frame in the first image crosses the border；
The absolute location information of the lamp plate frame is converted into picture element position information of the lamp plate frame in second image；
According to picture element position information of the lamp plate frame in second image and second image acquisition device
The image of traffic lights determines the second ROI region；
According to deep learning algorithm, detection and classification processing are carried out to second ROI region, obtain id pairs of the traffic lights
The state for the traffic lights answered.
6. according to the method described in claim 4, it is characterized in that, second image collector is set to short focus camera.
7. the method according to claim 1, wherein after the method further include:
Obtain vehicle current velocity information and acceleration information；
According to the current location of vehicle and the mark map, calculating vehicle is currently at a distance from stop line；
According to the velocity information, the acceleration information, the distance and the traffic lights state, calculate vehicle at crossing
State.
8. a kind of equipment, which is characterized in that the equipment includes memory and processor, and the memory is used to store program,
The processor requires any method of 1-7 for perform claim.
9. a kind of computer program product comprising instruction, which is characterized in that when the computer program product on computers
When operation, so that the computer perform claim requires any method of 1-7.
10. a kind of computer readable storage medium, which is characterized in that be stored with computer on the computer readable storage medium
Program, the computer program realize method as claimed in claim 1 to 7 when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910354808.2A CN109949594B (en) | 2019-04-29 | 2019-04-29 | Real-time traffic light identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910354808.2A CN109949594B (en) | 2019-04-29 | 2019-04-29 | Real-time traffic light identification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109949594A true CN109949594A (en) | 2019-06-28 |
| CN109949594B CN109949594B (en) | 2020-10-27 |
Family
ID=67016595
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910354808.2A Active CN109949594B (en) | 2019-04-29 | 2019-04-29 | Real-time traffic light identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109949594B (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110543814A (en) * | 2019-07-22 | 2019-12-06 | 华为技术有限公司 | Traffic light identification method and device |
| CN111444810A (en) * | 2020-03-23 | 2020-07-24 | 东软睿驰汽车技术（沈阳）有限公司 | Traffic light information identification method, device, equipment and storage medium |
| CN111582189A (en) * | 2020-05-11 | 2020-08-25 | 腾讯科技（深圳）有限公司 | Traffic signal lamp identification method and device, vehicle-mounted control terminal and motor vehicle |
| CN112183382A (en) * | 2020-09-30 | 2021-01-05 | 深兰人工智能(深圳)有限公司 | Unmanned traffic light detection and classification method and device |
| CN112327855A (en) * | 2020-11-11 | 2021-02-05 | 东软睿驰汽车技术（沈阳）有限公司 | Control method and device for automatic driving vehicle and electronic equipment |
| CN112614365A (en) * | 2020-12-14 | 2021-04-06 | 北京三快在线科技有限公司 | Electronic map processing method and device |
| CN112639813A (en) * | 2020-02-21 | 2021-04-09 | 华为技术有限公司 | Automatic driving control method, information processing method, device and system |
| CN112880692A (en) * | 2019-11-29 | 2021-06-01 | 北京市商汤科技开发有限公司 | Map data annotation method and device and storage medium |
| CN112991290A (en) * | 2021-03-10 | 2021-06-18 | 北京百度网讯科技有限公司 | Image stabilization method and device, road side equipment and cloud control platform |
| CN112991791A (en) * | 2019-12-13 | 2021-06-18 | 上海商汤临港智能科技有限公司 | Traffic information identification and intelligent driving method, device, equipment and storage medium |
| CN113177522A (en) * | 2021-05-24 | 2021-07-27 | 的卢技术有限公司 | Traffic light detection and identification method used in automatic driving scene |
| CN113178079A (en) * | 2021-04-06 | 2021-07-27 | 青岛以萨数据技术有限公司 | Marking system, method and storage medium for signal lamp and lane line |
| CN114299716A (en) * | 2021-12-27 | 2022-04-08 | 北京世纪高通科技有限公司 | Signal lamp time information correlation method and device, storage medium and equipment |
| CN114383594A (en) * | 2020-10-06 | 2022-04-22 | 丰田自动车株式会社 | Map generation device, map generation method, and computer program for map generation |
| CN114946201A (en) * | 2022-04-13 | 2022-08-26 | 北京小米移动软件有限公司 | Information transmission method and device and storage medium |
| CN115098606A (en) * | 2022-05-30 | 2022-09-23 | 九识智行(北京)科技有限公司 | Traffic light query method and device for unmanned vehicle, storage medium and equipment |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103680177A (en) * | 2013-12-03 | 2014-03-26 | 上海交通大学 | Intelligent vehicle speed prompting driving system based on mobile phone |
| CN105930819A (en) * | 2016-05-06 | 2016-09-07 | 西安交通大学 | System for real-time identifying urban traffic lights based on single eye vision and GPS integrated navigation system |
| CN106504554A (en) * | 2016-09-30 | 2017-03-15 | 乐视控股（北京）有限公司 | The method and device of identification traffic light status information |
| CN107618510A (en) * | 2016-07-13 | 2018-01-23 | 罗伯特·博世有限公司 | For the method and apparatus at least one driving parameters for changing vehicle during traveling |
| CN108305475A (en) * | 2017-03-06 | 2018-07-20 | 腾讯科技（深圳）有限公司 | A kind of traffic lights recognition methods and device |
| CN108706009A (en) * | 2017-03-31 | 2018-10-26 | 株式会社斯巴鲁 | The drive-control system of vehicle |
| CN109492507A (en) * | 2017-09-12 | 2019-03-19 | 百度在线网络技术（北京）有限公司 | The recognition methods and device of the traffic light status, computer equipment and readable medium |
-
2019
- 2019-04-29 CN CN201910354808.2A patent/CN109949594B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103680177A (en) * | 2013-12-03 | 2014-03-26 | 上海交通大学 | Intelligent vehicle speed prompting driving system based on mobile phone |
| CN105930819A (en) * | 2016-05-06 | 2016-09-07 | 西安交通大学 | System for real-time identifying urban traffic lights based on single eye vision and GPS integrated navigation system |
| CN107618510A (en) * | 2016-07-13 | 2018-01-23 | 罗伯特·博世有限公司 | For the method and apparatus at least one driving parameters for changing vehicle during traveling |
| CN106504554A (en) * | 2016-09-30 | 2017-03-15 | 乐视控股（北京）有限公司 | The method and device of identification traffic light status information |
| CN108305475A (en) * | 2017-03-06 | 2018-07-20 | 腾讯科技（深圳）有限公司 | A kind of traffic lights recognition methods and device |
| CN108706009A (en) * | 2017-03-31 | 2018-10-26 | 株式会社斯巴鲁 | The drive-control system of vehicle |
| CN109492507A (en) * | 2017-09-12 | 2019-03-19 | 百度在线网络技术（北京）有限公司 | The recognition methods and device of the traffic light status, computer equipment and readable medium |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110543814A (en) * | 2019-07-22 | 2019-12-06 | 华为技术有限公司 | Traffic light identification method and device |
| WO2021013193A1 (en) * | 2019-07-22 | 2021-01-28 | 华为技术有限公司 | Traffic light identification method and apparatus |
| CN110543814B (en) * | 2019-07-22 | 2022-05-10 | 华为技术有限公司 | Traffic light identification method and device |
| US11915492B2 (en) | 2019-07-22 | 2024-02-27 | Huawei Technologies Co., Ltd. | Traffic light recognition method and apparatus |
| CN112880692B (en) * | 2019-11-29 | 2024-03-22 | 北京市商汤科技开发有限公司 | Map data labeling method and device and storage medium |
| CN112880692A (en) * | 2019-11-29 | 2021-06-01 | 北京市商汤科技开发有限公司 | Map data annotation method and device and storage medium |
| CN112991791A (en) * | 2019-12-13 | 2021-06-18 | 上海商汤临港智能科技有限公司 | Traffic information identification and intelligent driving method, device, equipment and storage medium |
| CN112639813A (en) * | 2020-02-21 | 2021-04-09 | 华为技术有限公司 | Automatic driving control method, information processing method, device and system |
| WO2021164018A1 (en) * | 2020-02-21 | 2021-08-26 | 华为技术有限公司 | Automatic driving control method and apparatus, information processing method and apparatus, and system |
| CN111444810A (en) * | 2020-03-23 | 2020-07-24 | 东软睿驰汽车技术（沈阳）有限公司 | Traffic light information identification method, device, equipment and storage medium |
| CN111582189A (en) * | 2020-05-11 | 2020-08-25 | 腾讯科技（深圳）有限公司 | Traffic signal lamp identification method and device, vehicle-mounted control terminal and motor vehicle |
| CN112183382A (en) * | 2020-09-30 | 2021-01-05 | 深兰人工智能(深圳)有限公司 | Unmanned traffic light detection and classification method and device |
| CN114383594A (en) * | 2020-10-06 | 2022-04-22 | 丰田自动车株式会社 | Map generation device, map generation method, and computer program for map generation |
| CN112327855A (en) * | 2020-11-11 | 2021-02-05 | 东软睿驰汽车技术（沈阳）有限公司 | Control method and device for automatic driving vehicle and electronic equipment |
| CN112614365B (en) * | 2020-12-14 | 2022-07-15 | 北京三快在线科技有限公司 | Electronic map processing method and device |
| CN112614365A (en) * | 2020-12-14 | 2021-04-06 | 北京三快在线科技有限公司 | Electronic map processing method and device |
| CN112991290B (en) * | 2021-03-10 | 2023-12-05 | 阿波罗智联(北京)科技有限公司 | Image stabilizing method and device, road side equipment and cloud control platform |
| CN112991290A (en) * | 2021-03-10 | 2021-06-18 | 北京百度网讯科技有限公司 | Image stabilization method and device, road side equipment and cloud control platform |
| CN113178079A (en) * | 2021-04-06 | 2021-07-27 | 青岛以萨数据技术有限公司 | Marking system, method and storage medium for signal lamp and lane line |
| CN113178079B (en) * | 2021-04-06 | 2022-08-23 | 青岛以萨数据技术有限公司 | Marking system, method and storage medium for signal lamp and lane line |
| CN113177522A (en) * | 2021-05-24 | 2021-07-27 | 的卢技术有限公司 | Traffic light detection and identification method used in automatic driving scene |
| CN114299716A (en) * | 2021-12-27 | 2022-04-08 | 北京世纪高通科技有限公司 | Signal lamp time information correlation method and device, storage medium and equipment |
| CN114946201A (en) * | 2022-04-13 | 2022-08-26 | 北京小米移动软件有限公司 | Information transmission method and device and storage medium |
| CN114946201B (en) * | 2022-04-13 | 2023-11-28 | 北京小米移动软件有限公司 | Information transmission method and device and storage medium |
| CN115098606A (en) * | 2022-05-30 | 2022-09-23 | 九识智行(北京)科技有限公司 | Traffic light query method and device for unmanned vehicle, storage medium and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109949594B (en) | 2020-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109949594A (en) | Real-time traffic light recognition method | |
| JP7009716B2 (en) | Sparse map for autonomous vehicle navigation | |
| CN111291676B (en) | Lane line detection method and device based on laser radar point cloud and camera image fusion and chip | |
| CN109583415B (en) | Traffic light detection and identification method based on fusion of laser radar and camera | |
| CN112700470B (en) | Target detection and track extraction method based on traffic video stream | |
| EP3647734A1 (en) | Automatic generation of dimensionally reduced maps and spatiotemporal localization for navigation of a vehicle | |
| US20180273031A1 (en) | Travel Control Method and Travel Control Apparatus | |
| CN108896994A (en) | A kind of automatic driving vehicle localization method and equipment | |
| WO2021057134A1 (en) | Scenario identification method and computing device | |
| WO2020042348A1 (en) | Method for generating autonomous driving navigation map, system, vehicle-mounted terminal, and server | |
| CN109697860A (en) | Parking stall measure and tracking system and method and vehicle | |
| CN106503653A (en) | Area marking method, device and electronic equipment | |
| WO2020232648A1 (en) | Lane line detection method, electronic device and storage medium | |
| JP2020532800A (en) | Camera calibration systems and methods using traffic sign recognition, and computer-readable media | |
| CN110531376A (en) | Detection of obstacles and tracking for harbour automatic driving vehicle | |
| CN108594244B (en) | Obstacle recognition transfer learning method based on stereoscopic vision and laser radar | |
| CN109903574A (en) | The acquisition methods and device of crossing traffic information | |
| US20220355818A1 (en) | Method for a scene interpretation of an environment of a vehicle | |
| US20210300356A1 (en) | Vehicle uncertainty sharing | |
| US20230186646A1 (en) | System and method for detecting road and lane connections at intersections | |
| CN110765224A (en) | Processing method of electronic map, vehicle vision repositioning method and vehicle-mounted equipment | |
| CN113988197A (en) | Multi-camera and multi-laser radar based combined calibration and target fusion detection method | |
| CN117079238A (en) | Road edge detection method, device, equipment and storage medium | |
| US20230154203A1 (en) | Path planning method and system using the same | |
| Alrousan et al. | Multi-Sensor Fusion in Slow Lanes for Lane Keep Assist System |
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 | ||
| CP01 | Change in the name or title of a patent holder |
Address after: B4-006, maker Plaza, 338 East Street, Huilongguan town, Changping District, Beijing 100096 Patentee after: Beijing Idriverplus Technology Co.,Ltd. Address before: B4-006, maker Plaza, 338 East Street, Huilongguan town, Changping District, Beijing 100096 Patentee before: Beijing Idriverplus Technology Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder |