CN106485197A - Vehicle side image recognition method and device, car lamp error recognition detection and driving safety prediction method - Google Patents
Vehicle side image recognition method and device, car lamp error recognition detection and driving safety prediction method Download PDFInfo
- Publication number
- CN106485197A CN106485197A CN201610619785.XA CN201610619785A CN106485197A CN 106485197 A CN106485197 A CN 106485197A CN 201610619785 A CN201610619785 A CN 201610619785A CN 106485197 A CN106485197 A CN 106485197A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- identified
- image
- side image
- vehicular sideview
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/08—Detecting or categorising vehicles
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of vehicle side image recognition method and device, car lamp error recognition detection and driving safety prediction method, and the method comprising the steps of:What input camera head was absorbed comprises to identify the image of vehicle;Identify described in sorting out from the image of described input that vehicle is located at the image of described camera head left and right sides position;Vehicular sideview identifying processing is carried out to the described image sorting out, to have identified the side image information of vehicle described in obtaining;And, the described side image information having identified vehicle that output obtains.DAS (Driver Assistant System) more fully analyzes the mutual movement relation of target vehicle and other vehicles using the vehicle side image information that the method and device are provided, thus being greatly enhanced the safety of vehicle traveling.
Description
Technical field
The present invention relates to vehicle image identification field, more particularly, to a kind of vehicle image recognition methodss and device, Yi Zhongche
Lamp misrecognition detection method and a kind of vehicle driving safety Forecasting Methodology.
Background technology
By identifying the side image information of other vehicles obtained from vehicle side image, DAS (Driver Assistant System) can be made
Understand the transmits information of other vehicles by analyzing the direction of vehicular sideview so that DAS (Driver Assistant System) can be more comprehensively
Analysis target vehicle and other vehicles mutual movement relation, thus being greatly enhanced the safety of vehicle traveling.
Further, since vehicular sideview horizontal structure has certain similarity and in night vehicle with the horizontal structure of front face
The reflective similarity with its car car light of car light on side is so that the vehicle image based on vehicle structure, profile and car light identifies
Algorithm, often situation about misidentifying on vehicular sideview position.Therefore, identification vehicle side image also contributes to reduce
Vehicle image recognition methodss misrecognition in these cases.
The recognition methodss of existing identification vehicle side image, mainly pass through installing video camera, typing from car side
It is located at the vehicle side image of the other vehicles from car side.Then according to typing positioned at other vehicular sideviews from car side
Two side wheel images should be at the feature of same level position, to identify the wheel of vehicle being located at from car side, thus
Identification is located at the vehicle from car side.
But, by analysis, this vehicular sideview recognition methods, due to needing two side cars of vehicle to be identified
Wheel is located in same horizontal line, is only used in this way in the vehicle identification system installing video camera from car side.Cannot
It is front and side vehicle identification system offer vehicle side image identification function.
Content of the invention
Make the present invention with the forgoing problems and concerns in mind, and a kind of it is an object of the present invention to provide vehicle image
Recognition methodss and device, the vehicle side image information being provided using the method and device, DAS (Driver Assistant System) can be more comprehensively
Ground analysis target vehicle and the mutual movement relation of other vehicles, thus be greatly enhanced the safety of vehicle traveling.
Another mesh ground of the present invention is to provide a kind of vehicle image recognition methodss and device, using the method and device
The vehicle side image information providing, can reduce the misrecognition of vehicle image recognition methodss.
Further object is that providing a kind of car lamp error recognition detection method, it utilizes vehicle side image to believe
Breath detection car light misrecognition.
Further object is that providing a kind of vehicle driving safety Forecasting Methodology, it utilizes vehicle side image
Information prediction driving safety.
In order to realize the purpose of the present invention, according to a kind of vehicle side image recognition method of the present invention, including step:
What input camera head was absorbed comprises to identify the image of vehicle;
Identify described in sorting out from the image of described input that vehicle is located at described camera head left and right sides position
Image;
Vehicular sideview identifying processing is carried out to the described image sorting out, to have identified the side image of vehicle described in obtaining
Information;And
The described side image information having identified vehicle that output obtains.
In order to realize the purpose of the present invention, according to a kind of vehicle side image identifying device of the present invention, including:
Input block, for comprising of inputting that camera head absorbed identifying the image of vehicle;
Taxon, left for having identified that vehicle is located at described camera head described in sorting out from the image of described input
The image of right two side positions;
Identifying processing unit, for vehicular sideview identifying processing is carried out to the described image sorting out, with described in obtaining
The side image information of identification vehicle;And
Output unit, for having identified the side image information of vehicle described in output acquisition.
In order to realize the purpose of the present invention, according to a kind of car lamp error recognition detection method of the present invention, including step:
Input identified the image information of car light and camera head picked-up comprise identify the image of vehicle;
According to inputted comprise described in identified the image of vehicle, identified the side image coordinate of vehicle described in acquisition
Information;
Identify the image information of car light and the described side image coordinate information having identified vehicle according to described, determine position
The car light of identification in the described side image region having identified vehicle;And
The car light of identification exporting described determination is as misrecognition car light.
In order to realize the purpose of the present invention, according to a kind of vehicle driving safety Forecasting Methodology of the present invention, including step:
Comprising of inputting that the camera head picked-up being arranged on target vehicle obtains identify the image of vehicle and described
Target vehicle and the described relative velocity having identified vehicle;
Based on the image having identified vehicle described in the comprising of described input, described in acquisition, identify that the side image of vehicle is sat
Mark information;
According to the described side image coordinate information having identified vehicle, calculate the identification being located at described target vehicle rear
Vehicle and the relative angle of described target vehicle;
According to the relative angle of described calculating, determine and be located at described target vehicle rear and relatively may with described target vehicle
The vehicle of identification colliding;
That predicts described determination identifies that vehicle and described target vehicle collide the required time;And
Export the time identifying vehicle and described prediction of described determination.
Brief description
The flow chart that Fig. 1 shows the vehicle side image recognition method of one embodiment of the invention;
The flow chart that Fig. 2 shows the vehicle side image identifying processing of one embodiment of the invention:
Fig. 3 shows and contains the image being located at the vehicle from car dead astern;
Fig. 4 shows the image containing the vehicle near vanishing line;
Fig. 5 A shows the vehicle side image of camera head picked-up;
Fig. 5 B shows the side underbody shade and vehicular sideview edge that Filtering Processing obtains;
Fig. 5 C shows the side underbody shade and vehicular sideview edge that extraction obtains;
Fig. 6 shows the block diagram of the vehicle side image identifying device of one embodiment of the invention;
The flow chart that Fig. 7 shows the car lamp error recognition detection method of one embodiment of the invention;
Fig. 8 shows the vehicle driving safety Forecasting Methodology of one embodiment of the invention.
Specific embodiment
The flow chart that Fig. 1 shows the vehicle side image recognition method of one embodiment of the invention.As shown in figure 1, it is first
First, in step sl, input what such camera head picked-up such as television camera being arranged on target vehicle obtained
Comprise the image of other vehicles having identified.For the ease of description, the other vehicles hereinafter having identified this
Referred to as identify vehicle.
Then, in step s 2, according to identifying vehicle location in the picture, the image being inputted is divided into three
Class:(1) identified that vehicle is located at the image that the just front or rear orientation of target vehicle is put;(2) identified that vehicle is located near vanishing line
The image of position;And (3) have identified that vehicle is located at the image of side position about target vehicle.
Then, in step s3, it is located at the image of side position about target vehicle using identifying vehicle, to identifying car
Carry out vehicle side image identifying processing, to obtain the side image information having identified vehicle.Retouch in detail below in conjunction with Fig. 2
State the vehicle side image identifying processing of the present invention.
In the present embodiment, as shown in figure 3, because the vehicle positioned at target vehicle dead astern does not have vehicle side in the picture
Face, so the vehicle being pointed to target vehicle dead astern does not carry out vehicle side image identification, is pointed to target vehicle dead ahead
Vehicle same.Additionally, as shown in figure 4, area in the picture is less due to the vehicle near vanishing line, so vehicle
Side is almost invisible, if therefore such vehicle is carried out with vehicle side image identification, error is larger, based on this reason
It is not pointed to the vehicle near vanishing line in the present embodiment and carry out vehicle side image identification yet.
Finally, in step s 4, the side image information identifying vehicle that output obtains.
The flow chart that Fig. 2 shows the vehicle side image identifying processing of one embodiment of the invention.As shown in Fig. 2 it is first
First, in step S31, using existing various technology, the such as applying date is on 2 28th, 2006, applies for artificial Shenyang Neusoft
Software limited company and the Chinese patent of Ani Hadachi Co. Ltd, invention entitled " vehicle image dividing method and device "
Underbody shade disclosed in application determines method, carries out underbody shade filtering to the left and right sides identifying vehicle in the picture
Process, to obtain the side underbody shade that this has identified vehicle.
Then, in step s 32, utilize such existing edge extracting method such as Sobel, Canny, from image
In the enterprising driving in the direction of the side underbody shade of this acquisition lateral edge Filtering Processing, to obtain the car that this has identified vehicle
Lateral edge.
Then, in step S33, judge whether the side underbody shade of this acquisition and vehicular sideview edge meet following bar
Part:With this, the position at the side underbody shade of this acquisition and vehicular sideview edge has identified whether the position of vehicle is close, for example
Whether the positional distance between them has identified the 15% of vehicle width less than this;The side underbody shade of this acquisition and vehicle side
Whether the direction at face edge is close to parallel (for example, whether the differential seat angle in both directions is less than 3 degree);And, the side of this acquisition
Whether the positional distance at underbody shade and vehicular sideview edge meets has identified the 2/5 of vehicle overall height more than this and has known less than this
The 3/5 of other vehicle overall height.
Finally, if it is decided that the side underbody shade of this acquisition and vehicular sideview edge meet above-mentioned bar in step S34
Part, then extract side underbody shade and the vehicular sideview edge of this acquisition, and the side underbody shade according to this extraction and vehicle
The coordinate of lateral edge has identified the side image coordinate information of vehicle determining this.Wherein, this has identified the side view of vehicle
As coordinate information be this extraction side underbody shade and vehicular sideview edge coordinate union.
Fig. 5 A shows the vehicle side image of camera head picked-up, and Fig. 5 B shows according to above-mentioned steps S31 and S32 filter
Ripple processes the side underbody shade obtaining and vehicular sideview edge, and Fig. 5 C shows and extracts, according to above-mentioned steps S34, the side obtaining
Face underbody shade and vehicular sideview edge.
Fig. 6 shows the block diagram of the vehicle side image identifying device of one embodiment of the invention.As Fig. 6 institute
Show, vehicle side image identifying device 100 includes input block 102, be arranged on the camera head on target vehicle for input
What picked-up obtained comprises to identify the image of vehicle;Taxon 104, for described in sorting out from the image of described input
Identification vehicle is located at the image of described target vehicle left and right sides position;Identifying processing unit 106, for sorting out to described
Image carries out vehicle side image identifying processing, to have identified the side image information of vehicle described in obtaining;And, output unit
108, for having identified the side image information of vehicle described in exporting.
Wherein identifying processing unit 106 further includes:Shade obtains module 1060, in the described image sorting out
In the described left and right sides having identified vehicle is filtered process, to have identified the side underbody shade of vehicle described in obtaining;
Edge obtains module 1062, for carrying out vehicle side on the direction of described side underbody shade from the described image sorting out
Face edge filter is processed, to have identified the vehicular sideview edge of vehicle described in obtaining;Judge module 1064, for judging to be obtained
Side underbody shade and vehicular sideview edge whether meet following condition:The side underbody shade being obtained and vehicular sideview side
Whether the position of edge and the described positional distance having identified vehicle are less than the 15% of the described width having identified vehicle;Obtained
Whether the differential seat angle between both directions at side underbody shade and vehicular sideview edge is less than 3 degree;And, the side being obtained
The positional distance at underbody shade and vehicular sideview edge whether meet more than the described overall height having identified vehicle 2/5 and less than should
The 3/5 of overall height;And, determining module 1066, for when judging to connect fruit as being, according to described side underbody shade and vehicle side
The coordinate information at face edge, has identified the side image coordinate information of vehicle described in determination.
The application of the vehicle side image recognition method of the two basic change present invention is described below.
First application is the car lamp error recognition detection method of the vehicle side image recognition method with reference to the present invention.
Because vehicular sideview horizontal structure is closely similar with the horizontal structure of front face and anti-on night vehicular sideview
The light of the car light direct irradiation of the car light light penetrated and vehicle has similarity, therefore existing based on vehicle structure, profile
To identify the vehicle image recognizer of vehicle with car light, often the vehicular sideview of reflection car light light to be misidentified into vehicle.
However, by using the vehicle side image recognition method of the present invention, being capable of detecting when this misrecognition.
The flow chart that Fig. 7 shows the car lamp error recognition detection method of one embodiment of the invention.As shown in fig. 7, in step
In rapid S50, input the image information having identified car light and comprise with what the camera head picked-up being arranged on target vehicle obtained
The image of identification vehicle.
In step S52, above-mentioned step S1-S4 and S31- are executed to the image comprising to identify vehicle of this input
S34, to obtain the side image coordinate information having identified vehicle.
In step S54, according to the image information identifying car light being inputted and the side identifying vehicle being obtained
Face image coordinate information, determines the car light of identification being located in the side image region having identified vehicle.
In step S56, determined by output, identify car light as misrecognition car light.
Second application is the vehicle driving safety Forecasting Methodology of the vehicle side image recognition method with reference to the present invention.
Obtain the side image information of vehicle by identifying vehicle side image, DAS (Driver Assistant System) can be made to pass through point
The transmits information to understand other vehicles for the direction of analysis vehicular sideview is so that DAS (Driver Assistant System) can more fully be analyzed
Target vehicle and the mutual movement relation of other vehicles, thus be greatly enhanced the safety of vehicle traveling.
Fig. 8 shows the vehicle driving safety Forecasting Methodology of one embodiment of the invention.As shown in figure 8, in step S80
In, input that the camera head picked-up being arranged on target vehicle obtains comprise to identify the image of vehicle and target vehicle with
Identify the relative velocity of vehicle.
In step S82, above-mentioned step S1-S4 and S31- are executed to the image comprising to identify vehicle of this input
S34, to obtain the side image coordinate information having identified vehicle.
In step S84, according to the side image coordinate information identifying vehicle being obtained, calculate and be located at target vehicle
The relative angle identifying vehicle-to-target vehicle at rear.
In a step s 86, according to the relative angle being calculated, determine and be located at target vehicle rear and relatively can with target vehicle
The vehicle of identification that can collide.
In step S88, determined by prediction, identify that vehicle-to-target vehicle collides the required time.
In step S90, vehicle and the time predicted determined by output, are identified.
Although in the above embodiment of the present invention, so that camera head is arranged on target vehicle as a example to describe,
It is to the invention is not limited in this.It is true that in the present invention, camera head may be mounted to take the photograph on fixed object
Take the image of road conditions.
It will be appreciated by those skilled in the art that embodiment disclosed above be only intended to illustrate this law bright, and not
It is to limit the present invention.
It should also be appreciated by one skilled in the art that various method and device disclosed in this invention can be without departing from this
Make various modifications in the case of bright essence and deform, therefore protection scope of the present invention is come really by appending claims
Fixed.
Claims (10)
1. a kind of vehicle side image recognition method, including step:What input camera head was absorbed comprises to identify vehicle
Image;Identify described in sorting out from the image of described input that vehicle is located at the figure of described camera head left and right sides position
Picture;Vehicular sideview identifying processing is carried out to the described image sorting out, to have identified the side image information of vehicle described in obtaining;
And the described side image information having identified vehicle that output obtains.
2. vehicle side image recognition method as claimed in claim 1, wherein, described vehicular sideview identifying processing step enters one
Step includes:The side of vehicle has been identified from the described left and right sides having identified vehicle described in obtaining in the described image sorting out
Underbody shade;Know described in obtaining on the described side underbody shade direction having identified vehicle from the described image sorting out
The vehicular sideview edge of other vehicle;And identified the side underbody shade of vehicle and the coordinate at vehicular sideview edge according to described
Information, has identified the side image coordinate information of vehicle described in determination.
3. vehicle side image recognition method as claimed in claim 2, wherein, also includes step:Car has been identified described in judgement
Side underbody shade and both directions at vehicular sideview edge between differential seat angle whether meet predetermined condition;And if
Judged result is yes, then identified the side underbody shade of vehicle and the coordinate information at vehicular sideview edge according to described, determined
The described side image coordinate information having identified vehicle.
4. vehicle side image recognition method as claimed in claim 3, wherein, described predetermined condition is less than for described differential seat angle
3 degree.
5. vehicle side image recognition method as claimed in claim 2, wherein, also includes step:Car has been identified described in judgement
The positional distance at side underbody shade and vehicular sideview edge whether meet predetermined condition;And if it is judged that be
It is then to have identified the side underbody shade of vehicle and the coordinate information at vehicular sideview edge according to described, identified described in determination
The side image coordinate information of vehicle.
6. vehicle side image recognition method as claimed in claim 5, wherein, described predetermined condition is that described positional distance is big
In the described overall height having identified vehicle 2/5 and less than this overall height 3/5.
7. vehicle side image recognition method as claimed in claim 2, wherein, also includes step:Car has been identified described in judgement
Side underbody shade and vehicular sideview edge and the described positional distance having identified vehicle whether meet predetermined condition;And
If it is judged that being yes, then identify the side underbody shade of vehicle and the coordinate information at vehicular sideview edge according to described,
The side image coordinate information of vehicle has been identified described in determination.
8. vehicle side image recognition method as claimed in claim 7, wherein, described predetermined condition is that described positional distance is little
In the described width having identified vehicle 15%.
9. the vehicle side image recognition method as described in any one claim of claim 1-8, wherein, described shooting
Device is arranged on target vehicle or fixed object.
10. a kind of vehicle side image identifying device, including:Input block, comprises for input that camera head absorbed
The image of identification vehicle;Taxon, for having identified that vehicle is taken the photograph described in being located at described in sorting out from the image of described input
Image as device left and right sides position;Identifying processing unit, for carrying out vehicular sideview identification to the described image sorting out
Process, to have identified the side image information of vehicle described in obtaining;And output unit, for identifying described in output acquisition
The side image information of vehicle.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610619785.XA CN106485197A (en) | 2016-07-30 | 2016-07-30 | Vehicle side image recognition method and device, car lamp error recognition detection and driving safety prediction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610619785.XA CN106485197A (en) | 2016-07-30 | 2016-07-30 | Vehicle side image recognition method and device, car lamp error recognition detection and driving safety prediction method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106485197A true CN106485197A (en) | 2017-03-08 |
Family
ID=58273172
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610619785.XA Pending CN106485197A (en) | 2016-07-30 | 2016-07-30 | Vehicle side image recognition method and device, car lamp error recognition detection and driving safety prediction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106485197A (en) |
-
2016
- 2016-07-30 CN CN201610619785.XA patent/CN106485197A/en active Pending
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN101211408B (en) | Vehicle side image recognition method and apparatus, car lamp error identification detection and driving safety prediction method | |
Grubb et al. | 3D vision sensing for improved pedestrian safety | |
EP3296923B1 (en) | A method of detecting an overtaking vehicle, related processing system, overtaking vehicle detection system and vehicle | |
US8064643B2 (en) | Detecting and recognizing traffic signs | |
CN105825185B (en) | Vehicle collision avoidance method for early warning and device | |
EP1671216B1 (en) | Moving object detection using low illumination depth capable computer vision | |
KR101141874B1 (en) | Apparatus, Method for Dectecting Critical Areas and Pedestrian Detection Apparatus Using Same | |
CN102765365B (en) | Pedestrian detection method based on machine vision and pedestrian anti-collision warning system based on machine vision | |
EP3161507B1 (en) | Method for tracking a target vehicle approaching a motor vehicle by means of a camera system of the motor vehicle, camera system and motor vehicle | |
KR100969995B1 (en) | System of traffic conflict decision for signalized intersections using image processing technique | |
WO2016129403A1 (en) | Object detection device | |
CN106169244A (en) | The guidance information utilizing crossing recognition result provides device and method | |
CN104778444A (en) | Method for analyzing apparent characteristic of vehicle image in road scene | |
CN105488453A (en) | Detection identification method of no-seat-belt-fastening behavior of driver based on image processing | |
CN103548069B (en) | For the method and apparatus identifying possible colliding object | |
EP2833096B1 (en) | Method for determining a current distance and/or a current speed of a target object based on a reference point in a camera image, camera system and motor vehicle | |
CN110622504A (en) | Method and device for the spatially resolved detection of objects outside a vehicle by means of a sensor installed in the vehicle | |
US20150243169A1 (en) | Traffic lane situation determining device and method for determining traffic lane situation | |
CN106650730A (en) | Turn signal lamp detection method and system in car lane change process | |
KR101748646B1 (en) | Apparatus And Method Detecting Side Vehicle Using Camera | |
CN106557754A (en) | A kind of vehicle detection at night and state judging method | |
Cualain et al. | Multiple-camera lane departure warning system for the automotive environment | |
KR20140104516A (en) | Lane detection method and apparatus | |
Álvarez et al. | Perception advances in outdoor vehicle detection for automatic cruise control | |
CN106485197A (en) | Vehicle side image recognition method and device, car lamp error recognition detection and driving safety prediction method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170308 |
|
WD01 | Invention patent application deemed withdrawn after publication |