CN109229206A - The detection method and system of a kind of vehicle and its steering angle - Google Patents
The detection method and system of a kind of vehicle and its steering angle Download PDFInfo
- Publication number
- CN109229206A CN109229206A CN201810867378.XA CN201810867378A CN109229206A CN 109229206 A CN109229206 A CN 109229206A CN 201810867378 A CN201810867378 A CN 201810867378A CN 109229206 A CN109229206 A CN 109229206A
- Authority
- CN
- China
- Prior art keywords
- image
- vehicle
- edge line
- tire
- edge
- 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
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B62—LAND VEHICLES FOR TRAVELLING OTHERWISE THAN ON RAILS
- B62D—MOTOR VEHICLES; TRAILERS
- B62D15/00—Steering not otherwise provided for
- B62D15/02—Steering position indicators ; Steering position determination; Steering aids
- B62D15/021—Determination of steering angle
Landscapes
- Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Combustion & Propulsion (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses the detection methods and system of a kind of vehicle and its steering angle, wherein the image that the method passes through at least one steering tire of acquisition vehicle in real time；Edge detection is carried out to current described image, obtains the bianry image comprising the tyre rim；Detect the edge line of the tire in the bianry image；The steering angle and steering direction of current vehicle are obtained according to the edge line.As it can be seen that the present invention can in real time, directly obtain the steering angle of wheel after carrying out a series of processing by the image to steering tire, without changing the steering structure of vehicle, accuracy is good.
Description
Technical field
Unmanned vehicle of the present invention drives field, and in particular to the detection method and system of vehicle and its steering angle.
Background technique
In the actual moving process of unmanned vehicle, it is necessary to the steering angle Real-time Feedback of vehicle to control system, Cai Nengshi
The closed loop feedback control of existing unmanned vehicle, to improve the real-time and accuracy of control, it is ensured that the driving safety of unmanned vehicle.Vehicle
It all may cause vehicle in understeer or ovdersteering to break away, racing or overturning can occur when situation is serious, if energy
It is accurately obtained the real-time steering angle of vehicle, then the influence in combination with vehicle dynamic model analysis steering angle to vehicle driving, from
And promote the comprehensive performance of unmanned vehicle.Dissection test and non-disintegration are broadly divided into for the corner test detection of unmanned vehicle at this stage
Detect two major classes.
Application No. is 201711045769.5 patent disclose a kind of discontinuous test of electric automobile steering wheel corner and
Brake safe system, belongs to technical field of vehicle safety, including steering shaft and piezo-electric device, and steering shaft is filled by nose bar and piezoelectricity
Connection is set, piezo-electric device setting is in the fixed device of piezoelectricity, and the fixed device of piezoelectricity is arranged in shell, and steering shaft turns for steering wheel
To axis, it is fixed together with steering wheel；Nose bar includes the body of rod, idler wheel and pin shaft, and the body of rod is fixed in steering shaft, and idler wheel passes through pin
Axis is fixed on the body of rod, and idler wheel is rotated around pin shaft, and steering axes drive body of rod rotation, and the body of rod drives idler wheel rotation, idler wheel
It is contacted with piezo-electric device.The invention has the function of steering, display, automated steering braking etc., utilizes piezoelectric structure to collect steering and believes
Number, steering wheel angle is found out by control system accumulator, while vehicle safety is analyzed according to speed and angle relation, and control
Abs braking system processed improves safety.
Document " a kind of azimuthal new method of measurement Vehicular turn based on DSP " installs increment type on steering wheel for vehicle
Photoelectric encoder, accuracy in detection is higher, but this method changes the original structure of vehicle, is unfavorable for the upper road traveling of vehicle.
Malinen etc. estimates steering wheel angle " Fault detection of steering wheel by kinematical constraint
Sensor signal in an active front stering system ", and be used in the design of active front steering system,
The steering wheel angle that the algorithm obtains is estimated value, can monitor the abnormal operation of steering wheel, avoids the generation of major accident, but
The accuracy of thus obtained pivoted wheels on vehicle steering angle is not high, therefore cannot be used for the accurate control design case of system.
The state side that document " the corner of vehicle steering wheel identification based on method of inverse " is inputted with steering wheel for vehicle corner
Based on journey, the side acceleration response obtained using closed-loop simulation is inputted as inverse system, and the output of inverse system is original system
Input direction disk corner, this method identification steering wheel angle and test the steering wheel angle movement tendency of vehicle it is similar, but
Do not identify that the error of angle is analyzed to it in text.In each method for being directly realized by wheel turning angle detection, mainly
The steering angle detection of deflecting roller is realized using non-image sensor, such as " twin shaft turns to using platen rotating disk mechanism detection steering angle
The research of automotive wheel corner Auto-Test System ", but this method can not realize the detection to steering angle in normal vehicle operation.
Document " intelligent vehicle speed and the research of front-wheel steer deflection angle measuring systems " uses hardware angular displacement sensor method, but only provides
The testing result of 2 deflecting rollers, is only compared to each other, without will test by the experimental result in one corner direction
Value is compared with other methods.
It can be seen that existing unmanned vehicle steering angle detection, needs to change the steering structure of original automatic driving vehicle, cost
It is high and there are security risks；Or it is detected indirectly by steering wheel, but accuracy is not high.
Summary of the invention
Present invention generally provides the detection methods and system of a kind of vehicle and its steering angle, without changing the steering of vehicle
Structure detects steering angle by image recognition technology, improves the accuracy of detection.
According in a first aspect, provide a kind of detection method of vehicle steering angle in a kind of embodiment, include the following steps:
Image acquisition step obtains the image of at least one steering tire of vehicle in real time；
Edge detecting step carries out edge detection to current described image, obtains the two-value comprising the tyre rim
Image；
Edge line determines step, detects the edge line of the tire in the bianry image；
Angle detecting step obtains the steering angle and steering direction of current vehicle according to the edge line.
The detection method of the vehicle steering angle, wherein the side for detecting the tire in the bianry image
The step of edge line includes:
In the bianry image of present image, the parameter of the edge line detected according to previous frame image obtains region of interest
Domain；The area of the area-of-interest is less than the area of bianry image；
The area-of-interest is detected to obtain the edge line of tire.
The detection method of the vehicle steering angle, wherein the side for detecting the tire in the bianry image
The step of edge line includes:
Hough transform is carried out to each characteristic point at edge in the bianry image；Angle needed for Hough transform is equal
It is even to be divided into n equal portions；
One group of corresponding default Hough transform engine is all made of to the angle of each equal portions while carrying out Hough transform, with
Detect the edge line of the tire.
The detection method of the vehicle steering angle, wherein described real-time at least one deflecting roller wheel for obtaining vehicle
The step of image of tire includes: that the imaging sensor being mounted on the left side mirror of vehicle or on right side mirror acquires correspondence in real time
Steering tire image.
According to second aspect, a kind of detection system of vehicle steering angle is provided in a kind of embodiment, comprising:
Image capture module, the image of at least one steering tire for obtaining vehicle in real time；
Edge detection unit is obtained for carrying out edge detection to current described image comprising the tyre rim
Bianry image；
Edge line determination unit, for detecting the edge line of the tire in the bianry image；
Angle detection unit, for obtaining the steering angle and steering direction of current vehicle according to the edge line.
The detection system of the vehicle steering angle, wherein the edge line determination unit is specifically used for:
In the bianry image of present image, the parameter of the edge line detected according to previous frame image obtains region of interest
Domain；The area of the area-of-interest is less than the area of bianry image；
The area-of-interest is detected to obtain the edge line of tire.
The detection system of the vehicle steering angle, wherein the edge line determination unit is specifically used for:
Hough transform is carried out to each characteristic point at edge in the bianry image；Angle needed for Hough transform is equal
It is even to be divided into n equal portions；
One group of corresponding default Hough transform engine is all made of to the angle of each equal portions while carrying out Hough transform, with
Detect the edge line of the tire.
The detection system of the vehicle steering angle, wherein further include being mounted on the left side mirror of vehicle or the right side is fallen
Imaging sensor on vehicle mirror；
Described image sensor, for acquiring the image of corresponding steering tire in real time.
According to the third aspect, a kind of vehicle is provided in a kind of embodiment, comprising:
Memory, for storing program；
Processor, for the program by executing the memory storage to realize method as described above.
According to fourth aspect, a kind of computer readable storage medium, including program, described program are provided in a kind of embodiment
It can be executed by processor to realize method as described above.
According to the vehicle of above-described embodiment and its detection method and system of steering angle, wherein the method passes through reality
When obtain vehicle at least one steering tire image；Edge detection is carried out to current described image, is obtained comprising institute
State the bianry image of tyre rim；Detect the edge line of the tire in the bianry image；Worked as according to the edge line
The steering angle and steering direction of vehicle in front.As it can be seen that after the present invention carries out a series of processing by the image to steering tire
The steering angle that can in real time, directly obtain wheel, without changing the steering structure of vehicle, accuracy is good.
Detailed description of the invention
Fig. 1 is that the detection system of vehicle steering angle provided by the invention unifies the structural block diagram of embodiment；
Fig. 2 is imaging sensor schematic view of the mounting position in the detection system of vehicle steering angle provided by the invention；
Fig. 3 is the structural block diagram of another embodiment of detection system of vehicle steering angle provided by the invention；
Fig. 4 is the schematic diagram of convolutional coding structure in the detection system of vehicle steering angle provided by the invention；
Fig. 5 is the flow diagram of Hough transform in the detection system of vehicle steering angle provided by the invention；
Fig. 6 is the rotationally-varying schematic diagram of circle in the detection system of vehicle steering angle provided by the invention；
Fig. 7 is the flow chart of the detection method of vehicle steering angle provided by the invention.
Specific embodiment
Below by specific embodiment combination attached drawing, invention is further described in detail.Wherein different embodiments
Middle similar component uses associated similar element numbers.In the following embodiments, many datail descriptions be in order to
The application is better understood.However, those skilled in the art can recognize without lifting an eyebrow, part of feature
It is dispensed, or can be substituted by other elements, material, method in varied situations.In some cases, this Shen
Please it is relevant it is some operation there is no in the description show or describe, this is the core in order to avoid the application by mistake
More descriptions are flooded, and to those skilled in the art, these relevant operations, which are described in detail, not to be necessary, they
Relevant operation can be completely understood according to the general technology knowledge of description and this field in specification.
It is formed respectively in addition, feature described in this description, operation or feature can combine in any suitable way
Kind embodiment.Meanwhile each step in method description or movement can also can be aobvious and easy according to those skilled in the art institute
The mode carry out sequence exchange or adjustment seen.Therefore, the various sequences in the description and the appended drawings are intended merely to clearly describe a certain
A embodiment is not meant to be necessary sequence, and wherein some sequentially must comply with unless otherwise indicated.
It is herein component institute serialization number itself, such as " first ", " second " etc., is only used for distinguishing described object,
Without any sequence or art-recognized meanings.And " connection ", " connection " described in the application, unless otherwise instructed, include directly and
It is indirectly connected with (connection).
Vehicle provided by the invention, as shown in Figure 1, including the detection system of vehicle steering angle, which includes: at least
One imaging sensor 1, image capture module 31 and processing module 32.
Imaging sensor 1, for acquiring the image of corresponding steering tire in real time.
Image capture module 31, the image exported for real-time image acquisition sensor 1.
Processing module 32, the image for currently obtaining to image capture module 31 carry out edge detection, obtain comprising wheel
The bianry image of tyre bead edge；Detect the edge line of the tire in bianry image；Current vehicle is obtained according to the edge line
Steering angle and steering direction.
As it can be seen that the present invention in real time, can be obtained directly after carrying out a series of processing by the image to steering tire
The steering angle of wheel, without changing the steering structure of vehicle, accuracy is good.System provided by the invention, due to real-time and
Accuracy is good, is very suitable to the steering angle of detection unmanned vehicle, greatly improves the control precision and safety of unmanned vehicle.
Imaging sensor 1 it is settable one or two, in the present embodiment, as shown in Figures 2 and 3, imaging sensor 1 be arranged
Two, be separately mounted on the left side mirror of vehicle and right side mirror on.Left-side images sensor is towards the near front wheel, image right
Sensor is towards off-front wheel.It takes such mounting means without disassembling body structure, can more intuitively obtain the real-time of front-wheel
Image.And the installation site of imaging sensor 1 is higher, can effectively ensure to avoid the influence of dust from land surface and rainwater
The clarity and stability of acquired image improve the precision of steering angle detection.
In the present embodiment, imaging sensor 1 uses high speed imaging sensor, for example, using high speed camera.Its processing speed
Fastly, high frame frequency shooting may be implemented, Image Information Processing and transmittability are strong, and support random window read functions, improve
Flexibility.
When high speed imaging sensor is chosen, cmos image sensor is selected, is had relative to traditional ccd sensor
The significant advantages such as processing speed is fast, low in energy consumption, degree of integration is high, strong antijamming capability.
Imaging sensor 1 is conducive to letter so that the compatibility of imaging sensor 1 is stronger using one group of working power power supply
Change high speed camera power supply architecture, lengthens working hours.Cmos image sensor size is small, integrated level is high, without configuring image letter
The external circuits such as number amplifier save space and simultaneously simplify high speed camera circuit structure, are advantageously implemented the miniaturization of high speed camera,
It is easily mounted on the side mirror of vehicle；External circuit is reduced, and integrated level height also can be improved the stability of high speed camera.
Referring to Fig. 3, the system also includes signal converting subcard 2, FPGA board 3 and vehicle-mounted computer 4.Described image is adopted
Collect module 31 and processing module 32 is arranged on FPGA board 3.Control module 33, communication module are additionally provided on FPGA board 3
34 and data transmission module 35.
The high speed imaging sensor 1 being mounted on left and right vehicle wheel side mirror passes through signal converting subcard 2 and FPGA board 3
Image capture module 31 is connected.FPGA board 3 is transmitted after handling the Data Data of high speed imaging sensor 1 by data
Module 35 is sent to vehicle-mounted computer 4.
Optical system (for example, camera lens) collected optical signal is converted to electric signal by high speed imaging sensor 1, and in piece
It is interior analog signal is handled after in the form of digital signal export.
Control module 33 is that imaging sensor 1 provides driver' s timing, when controlling operating mode, the integral of imaging sensor 1
Between, frame frequency etc., and coordinate the work of whole system.
Processing module 32 is sent to data after handling the collected digital image data of image capture module 31
Transmission module 35.
Data transmission module 35 is based on data and transmits chip, and digital image data is transmitted according to specified communications protocol
To the master control system of vehicle-mounted computer 4.The present invention makes full use of the performance and advantage of data transmission chip, establishes real with master control system
When communicate, realize the real-time Transmission of digital image data.
Communication module 34 realizes that imaging system control instruction receives and working condition feedback, the responsible master control system that receives issue
Control command.Control command is used to control the operating mode of control module 33 Yu processing module 32.Communication module 34 is also used to
The working condition of control module 33 and processing module 32 etc. is fed back into master control system, for monitoring the detection system of steering angle
Operating status.
Wherein, control module 33 and processing module 32 are responsible in detection system the co-ordination between modules, control figure
It is the nucleus module of detection system as working sensor.
Further, processing module 32 includes that edge detection unit 321, edge line determination unit 322 and angle detection are single
Member 323.
Edge detection unit 321, the image for currently obtaining to image capture module 31 carry out edge detection, are wrapped
Bianry image containing the tyre rim.Edge detection is to detect one of the important step of tyre rim line, the direct shadow of quality
Ring the effect of subsequent algorithm.By being compared analysis to current multiple edge detection algorithms, the present invention chooses traditional canny
Algorithm (Canny edge detection algorithm) or the window convolution Sobel algorithm (sobel algorithm) of suitable FPGA operation carry out edge
It detects, both the preferable edge Binarization methods of effect.In the present embodiment, due to using FPGA board, therefore use suitable
Close the window convolution Sobel algorithm of FPGA operation.
As shown in figure 4, carrying out operation using the time parallel and spatial parallelism of pipeline organization, very in FPGA board
It is suitable for realizing the algorithm based on convolutional coding structure, to improve the operation efficiency of edge detection.
Edge detection unit 321 includes line shift register and trigger；The image that image capture module 31 obtains is to scheme
As the form of data is as successively flowing water enters in FPGA (in edge detection unit 321) system clock, pass through row shift LD
Device and trigger can simultaneously obtain the data in image in adjacent m × m window, and handle it, i.e. progress edge two
Value processing.Therefore, in FPGA realize convolutional coding structure algorithm when can while not saving entire image, while access according to while
It performs corresponding processing, therefore, very high efficiency is had based on the filtering algorithm of convolutional coding structure to realization such as Sobel algorithm.
Wherein, m is the integer more than or equal to 1, and m is 3 in the present embodiment.
In view of tyre rim has inside edge and outer ledge, if it is possible to while being examined using the information of dual edge
The edge of measuring wheel tire, and carry out the accuracy rate that Model Matching will improve angle detection.Therefore, if two edge detections can be used
Template detects inner edge and the outer rim of tire respectively, then matches to them, can greatly improve the correct of edge detection
Rate.And traditional processor is serial computing structure, the increase of edge detection template quantity can cause the linear of operation time
Increase, to system it is real-time handle more stringent requirements are proposed.As it can be seen that the present invention utilizes multiple and different window convolution templates
(i.e. above-mentioned edge detection template) detects inside edge and the outer ledge of tire simultaneously, then by carrying out to two edges
PARALLEL MATCHING obtains two edges of tire.Since the core of its algorithm is all based on window convolution technology, edge inspection
Method of determining and calculating can real-time implementation.
After the processing of edge detection unit 321 obtains the bianry image comprising outer ledge in tire, need to identify interior outside
Edge line, therefore the edge line of the tire in bianry image is detected by edge line determination unit 322.For example, dividing (planning)
The area-of-interest of edge line in bianry image, and then area-of-interest is detected to obtain the edge line of tire.It is interested
The area in region is less than the area of bianry image.The area-of-interest for dividing edge line in bianry image, can be according to tyre beads
The body contour of edge and vehicle divides, and rejects the background other than tyre rim line.It can also be detected according to previous frame image
The parameter of edge line carry out interested area division.By interested area division, subsequent calculation amount is greatly reduced, is mentioned
High efficiency.
In the present embodiment, edge line determination unit 322 is detected in the bianry image of present image according to previous frame image
The parameter of edge line out obtains area-of-interest.The edge line parameter of current frame image can be joined with previous frame image border line
Range is limited based on number to carry out obtaining area-of-interest, and edge line can be rapidly detected in area-of-interest.Edge
The parameter of line can be the coordinate of edge line.Edge line based on intra-frame trunk is detected, and image data is not only reduced
Operand, reduces the execution time of algorithm, and improves the robustness of algorithm.
The present embodiment is illustrated by taking three frame images as an example:
(1) the tyre rim line parameter of former frame figure is (ρ 1, i-1, θ 1, i-1) and (ρ 2, i-1, θ 2, i-1), to preceding
One frame parameter, which carries out constraint, can obtain the area-of-interest of present frame；
(2) area-of-interest carry out Hough transform (Hough transformation) obtain present frame tyre rim line parameter (ρ 1,
I, θ 1, i) and (ρ 2, i, θ 2, i), obtain the area-of-interest that next frame image can be determined after the parameter of present frame；Hough
Transformation is to identify one of the basic skills of geometry in image procossing from image；
(3) Hough transform is carried out in the area-of-interest of next frame image and obtain tyre rim line parameter, repeat (1)
(2), the tyre rim line tracking of video flowing can be obtained.
After edge line determination unit 322 obtains area-of-interest, subregion is carried out to area-of-interest；Subregion is known
Not, the set of curves in each area is extracted；The geometrical characteristic of each subregion inner curve cluster is analyzed to determine the curved section for belonging to edge line；It will
Each subregion is divided into medial area and outside area.The curved section of each edge line, which connects, in medial area just knows from bianry image
The edge line being clipped on the inside of tire, the curved section of each edge line, which connects, in outside area just recognizes from bianry image
Edge line on the outside of tire.Wherein, the geometrical characteristic of each subregion inner curve cluster is analyzed to determine the curved section for belonging to edge line,
Hough transform can be used to each subregion inner curve cluster, so that it is determined that belonging to the curved section of edge line.
The advantage and characteristic of the present embodiment combination FPGA parallel processing, improves traditional Hough transform, becomes
Change of current journey is as shown in figure 5, likewise, Hough transform is using spatially parallel processing and time upper pipeline processes.In space
On, in conjunction with the characteristics of the operation simultaneously of FPGA multicore, according to the edge (edge image) of bianry image, store each subregion inner curve cluster
The location information of characteristic point on (set of curves indicates all kinds of edges), according to these characteristic points, the required angle of Hough transform
Uniformly it is divided into n equal portions, individually uses one group of Hough transform engine to carry out conversion process the angle of each equal portions, to detect
Belong to the curved section of edge line, to improve its working efficiency.The corresponding one group of Hough transform engine of one equal portions, Hough transform
Engine is preset.As shown in figure 5, on n route, each group of Hough transform engine pipelining in time makes
The calculating each time of Hough transform can all be completed within a clock cycle.To greatly improve the effect of Hough transform
Rate.Wherein, n is the integer greater than 1.
The edge line of tire is usually ellipse, and obtaining edge line can be obtained corresponding elliptical focus.Become to reduce Hough
The memory space changed sets the origin of current Hough transform to the focus of the edge line of previous frame image, becomes Hough
The memory space changed greatly reduces.
Angle detection unit 323, for obtaining the steering angle and steering direction of current vehicle according to the edge line.Tool
Body, angle detection unit 323 carries out curve fitting to multiple points (such as characteristic point) of the curved section of each subregion, is intended
The curved section of the fitting of tire medial area and outside area is reconstructed in the curved section of conjunction, obtain edge line on the inside of tire and
The edge line in outside；The steering angle and steering direction of current vehicle are obtained according to the edge line that reconstruct obtains.For example, according to circle
Rotation angle detection method, to calculate preliminary steering angle and steering direction；Also to utilize convenient for eliminating noise and exceptional value
The continuity features of vehicle movement carry out Effective judgement to data, to obtain steering angle and the steering of current vehicle.
For example, steering angle variation is continuous when normal vehicle operation, and the time interval of adjacent two field pictures is fixed value, therefore adjacent
Significantly jumping should not occur in the steering angle of two field pictures.It, can according to the variable quantity of continuous several groups of consecutive frame steering angles
The steering trend for knowing vehicle is that can determine that the validity of steering using the steering trend.Certainly, angle detection unit 323 may be used also
With the continuous several groups of consecutive frame edge lines parameter (edge line coordinates) obtained according to edge line determination unit 322, the company of being calculated
The variable quantity for continuing several groups of consecutive frame edge line parameters, compares the size of these variable quantities Yu preset threshold value, if variable quantity exceeds
Threshold value illustrates that steering angle is abnormal, then it is believed that corresponding steering angle is invalid.
In the process of moving, front-wheel (deflecting roller) will appear camber angle and toe-in angle to vehicle, the steering relative to deflecting roller
Angle, the general very little of camber angle can be ignored, it can think that deflecting roller is only angled on the direction perpendicular to ground.Deflecting roller
Outer edge presentation is approximately round feature, and thus by deflecting roller circle substitution, therefore the present invention is using the rotation angle detection side of circle
Method realizes the detection to steering angle.
Specifically, angle detection unit 323 establishes rectangular coordinate system as shown in FIG. 6, tyre rim line is circle,
Steering angle when initial position is 0 °, and the tyre rim line seen in a coordinate system at this time is a straight line (perspective of edge line
Figure is straight line), i.e. vertical line in Fig. 6.For deflecting roller once steering, then the edge line seen in coordinate system becomes oval (side
The perspective view of edge line is ellipse).With the increase of steering angle, the elliptical short axle that perspective view is seen constantly increases, and long axis is protected
It holds constant.When steering angle increases to 90 °, elliptical long axis is equal and rounded with short axle.It will be appreciated from fig. 6 that rotation angle
Calculation formula are as follows:
α=arcsina/b (1)；
In formula: a is that the edge line that edge line determination unit 322 obtains has an X-rayed the elliptical short axle seen in a coordinate system；b
For the radius for the edge line that edge line determination unit 322 obtains, i.e., the described elliptical long axis.Known elliptic equation is
In formula: x0、y0Abscissa value, ordinate value for the oval center of circle.Formula (3) is represented by
Wherein,And A > 0, D > 0
As it can be seen that the value of a, b, the rotation angle of edge line can be acquired if elliptic parameter A, B, C, D in formula (4) can be obtained
Degree can be obtained by formula (1).Therefore, angle detection unit 323 carries out curve fitting to multiple points of the curved section of each subregion,
Equation of the curved section in the coordinate system is obtained, the curved section of fitting is reconstructed, to obtain inside edge line, outside
Edge line is located at the elliptical equation formed in coordinate system to get to elliptic parameter A, B, C, D in formula (4), is obtained according to formula (1)
To steering angle.
The present invention realizes the detection of vehicle steering angle using imaging sensor, and sensor is easily installed, not to vehicle knot
It is configured to destroy and accuracy in detection is high, the detection to steering angle can be realized in normal vehicle operation.In order to guarantee to detect
Accuracy, consider be easily installed on the basis of, the selection of the installation site, angle of high speed imaging sensor, should with obtain
Characteristics of image protrude be conducive to be identified as principle to carry out, in favor of the extraction to characteristics of image.
Sensor carries out contour edge extraction to collected deflecting roller image, and carries out curve matching, again to edge point
Structure；Then according to round rotation angle detection method, to calculate preliminary corner value and steering；Finally utilize the continuity of vehicle movement
Feature carries out Effective judgement to data, to obtain steering angle and the steering of current vehicle.
The detection system of the vehicle steering angle provided based on the above embodiment, the present invention also provides a kind of vehicle steering angles
The detection method of degree, as shown in fig. 7, comprises following steps:
Image acquisition step 10 obtains the image of at least one steering tire of vehicle in real time.For example, being mounted on vehicle
Left side mirror on or right side mirror on imaging sensor acquire the image of corresponding steering tire in real time；By Image Acquisition
The image of module real-time image acquisition sensor output.
Edge detecting step 20 carries out edge detection to current described image, obtains two comprising the tyre rim
It is worth image.The window convolution Sobel algorithm for choosing traditional canny algorithm or suitable FPGA operation carries out edge detection, the two
It is all the preferable edge Binarization methods of effect.In the present embodiment, due to using FPGA board, therefore uses and be suitble to FPGA operation
Window convolution Sobel algorithm.The image that image capture module obtains is in the form of image data as system clock successively flows
Water enters in FPGA (in edge detection unit), can be simultaneously by the line shift register and trigger of edge detection unit
The data in image in adjacent m × m window are obtained, and it is handled, is i.e. progress edge binary conversion treatment.Therefore, exist
In FPGA realize convolutional coding structure algorithm when can while not saving entire image, while access according to while perform corresponding processing,
Therefore, very high efficiency is had based on the filtering algorithm of convolutional coding structure to realization such as Sobel algorithm.Wherein, m be more than or equal to
1 integer, m is 3 in the present embodiment.
In view of tyre rim has inside edge and outer ledge, if it is possible to while being examined using the information of dual edge
The edge of measuring wheel tire, and carry out the accuracy rate that Model Matching will improve angle detection.Therefore, if two edge detections can be used
Template detects inner edge and the outer rim of tire respectively, then matches to them, can greatly improve the correct of edge detection
Rate.And traditional processor is serial computing structure, the increase of edge detection template quantity can cause the linear of operation time
Increase, to system it is real-time handle more stringent requirements are proposed.As it can be seen that the present invention utilizes multiple and different window convolution templates
(i.e. above-mentioned edge detection template) detects inside edge and the outer ledge of tire simultaneously, then by carrying out to two edges
PARALLEL MATCHING obtains two edges of tire.Since the core of its algorithm is all based on window convolution technology, edge inspection
Method of determining and calculating can real-time implementation.
Edge line determines step 30, detects the edge line of the tire in the bianry image.For example, dividing (planning) two
It is worth the area-of-interest of edge line in image, and then area-of-interest is detected to obtain the edge line of tire.Region of interest
The area in domain is less than the area of bianry image.The area-of-interest for dividing edge line in bianry image, can be according to tyre rim
The body contour of line and vehicle divides, and rejects the background other than tyre rim line.It can also be detected according to previous frame image
The parameter of edge line carry out interested area division.By interested area division, subsequent calculation amount is greatly reduced, is mentioned
High efficiency.
In the present embodiment, in the bianry image of present image, according to the parameter for the edge line that previous frame image detects
Obtain area-of-interest.The edge line parameter of current frame image can limit range based on the line parameter of previous frame image border
It carries out obtaining area-of-interest, edge line can be rapidly detected in area-of-interest.The parameter of edge line can be side
The coordinate of edge line.Edge line based on intra-frame trunk is detected, and is not only reduced the operand of image data, is reduced algorithm
The execution time, and improve the robustness of algorithm.
The present embodiment is illustrated by taking three frame images as an example:
(1) the tyre rim line parameter of former frame figure is (ρ 1, i-1, θ 1, i-1) and (ρ 2, i-1, θ 2, i-1), to preceding
One frame parameter, which carries out constraint, can obtain the area-of-interest of present frame；
(2) Hough transform is carried out in area-of-interest obtain the tyre rim line parameter (ρ 1, i, θ 1, i) and (ρ of present frame
2, i, θ 2, i), the parameter for obtaining present frame can determine the area-of-interest of next frame image later；
(3) Hough transform is carried out in the area-of-interest of next frame image and obtain tyre rim line parameter, repeat (1)
(2), the tyre rim line tracking of video flowing can be obtained.
After obtaining area-of-interest, subregion is carried out to area-of-interest；Subregion is identified, the curve in each area is extracted
Cluster；The geometrical characteristic of each subregion inner curve cluster is analyzed to determine the curved section for belonging to edge line；Each subregion is divided into medial area
And outside area.The curved section of each edge line connects just from the side recognized on the inside of tire in bianry image in medial area
Edge line, the curved section of each edge line connects just from the edge recognized on the outside of tire in bianry image in outside area
Line.Wherein, the geometrical characteristic of each subregion inner curve cluster is analyzed to determine the curved section for belonging to edge line, each subregion can be turned inward
Line cluster uses Hough transform, so that it is determined that belonging to the curved section of edge line.
The advantage and characteristic of the present embodiment combination FPGA parallel processing, improves traditional Hough transform, becomes
Change of current journey is as shown in figure 5, likewise, Hough transform is using spatially parallel processing and time upper pipeline processes.In space
On, in conjunction with the characteristics of the operation simultaneously of FPGA multicore, according to the edge (edge image) of bianry image, store each subregion inner curve cluster
The location information of characteristic point on (set of curves indicates all kinds of edges), according to these characteristic points, the required angle of Hough transform
Uniformly it is divided into n equal portions, individually uses one group of Hough transform engine to carry out conversion process the angle of each equal portions, to detect
Belong to the curved section of edge line, to improve its working efficiency.The corresponding one group of Hough transform engine of one equal portions, Hough transform
Engine is preset.As shown in figure 5, on n route, each group of Hough transform engine pipelining in time makes
The calculating each time of Hough transform can all be completed within a clock cycle.To greatly improve the effect of Hough transform
Rate.Wherein, n is the integer greater than 1.
The edge line of tire is usually ellipse, and obtaining edge line can be obtained corresponding elliptical focus.Become to reduce Hough
The memory space changed sets the origin of current Hough transform to the focus of the edge line of previous frame image, becomes Hough
The memory space changed greatly reduces.
Angle detecting step obtains the steering angle and steering direction of current vehicle according to the edge line.Specifically, angle
Degree detection unit 323 carries out curve fitting to multiple points (such as characteristic point) of the curved section of each subregion, the song being fitted
The curved section of the fitting of tire medial area and outside area is reconstructed in line segment, obtains edge line on the inside of tire and outside
Edge line；The steering angle and steering direction of current vehicle are obtained according to the edge line that reconstruct obtains.For example, according to round rotation
Angle detection method, to calculate preliminary steering angle and steering direction；Also to be transported using vehicle convenient for eliminating noise and exceptional value
Dynamic continuity features carry out Effective judgement to data, to obtain steering angle and the steering of current vehicle.For example, vehicle
Steering angle variation is continuous when normally travel, and the time interval of adjacent two field pictures is fixed value, therefore adjacent two frames figure
Significantly jumping should not occur in the steering angle of picture.According to the variable quantity of continuous several groups of consecutive frame steering angles, it is known that vehicle
Steering trend, be that can determine that the validity of steering using the steering trend.It is, of course, also possible to determine that step obtains according to edge line
The change of continuous several groups of consecutive frame edge line parameters is calculated in the continuous several groups of consecutive frame edge lines parameter (edge line coordinates) arrived
Change amount compares the size of these variable quantities Yu preset threshold value, if variable quantity exceeds threshold value, illustrates that steering angle is abnormal, then
It is believed that corresponding steering angle is invalid.
In the process of moving, front-wheel (deflecting roller) will appear camber angle and toe-in angle to vehicle, the steering relative to deflecting roller
Angle, the general very little of camber angle can be ignored, it can think that deflecting roller is only angled on the direction perpendicular to ground.Deflecting roller
Outer edge presentation is approximately round feature, and thus by deflecting roller circle substitution, therefore the present invention is using the rotation angle detection side of circle
Method realizes the detection to steering angle.
Specifically, establishing rectangular coordinate system as shown in FIG. 6, tyre rim line is circle, is turned in initial position
It is 0 ° to angle, the tyre rim line seen in a coordinate system at this time is a straight line (perspective view of edge line is straight line), i.e. Fig. 6
In vertical line.Deflecting roller is once steering, then the edge line seen in coordinate system becomes ellipse, and (perspective view of edge line is ellipse
Circle).With the increase of steering angle, the elliptical short axle that perspective view is seen constantly increases, and long axis remains unchanged.Work as steering angle
When increasing to 90 °, elliptical long axis is equal and rounded with short axle.It will be appreciated from fig. 6 that the calculation formula of rotation angle are as follows:
α=arcsina/b (1)；
In formula: a is that the edge line that edge line determination unit 322 obtains has an X-rayed the elliptical short axle seen in a coordinate system；b
For the radius for the edge line that edge line determination unit 322 obtains, i.e., the described elliptical long axis.Known elliptic equation is
In formula: x0、y0Abscissa value, ordinate value for the oval center of circle.Formula (3) is represented by
Wherein,And A > 0, D > 0
As it can be seen that the value of a, b, the rotation angle of edge line can be acquired if elliptic parameter A, B, C, D in formula (4) can be obtained
Degree can be obtained by formula (1).Therefore, angle detection unit carries out curve fitting to multiple points of the curved section of each subregion, obtains
To equation of the curved section in the coordinate system, the curved section of fitting is reconstructed, to obtain inside edge line, outer side edges
Edge line is located at the elliptical equation formed in coordinate system to get to elliptic parameter A, B, C, D in formula (4), is obtained according to formula (1)
Steering angle.
As it can be seen that method provided by the invention, just can accurately obtain steering angle without being changed to vehicle original structure
Degree.The scheme of vehicle actual steering angle is obtained by direct detection direction disk corner compared to existing, the accuracy of this method and
Real-time is more preferable.CMOS high speed imaging sensor+the FPGA architecture taken, reduce power consumption, simplified structure improve processing speed
Degree and anti-interference.
Since the detailed process and effect of method have elaborated in the above system embodiment, details are not described herein.
It will be understood by those skilled in the art that all or part of function of various methods can pass through in above embodiment
The mode of hardware is realized, can also be realized by way of computer program.When function all or part of in above embodiment
When being realized by way of computer program, which be can be stored in a computer readable storage medium, and storage medium can
To include: read-only memory, random access memory, disk, CD, hard disk etc., it is above-mentioned to realize which is executed by computer
Function.For example, program is stored in the memory of equipment, when executing program in memory by processor, can be realized
State all or part of function.In addition, when function all or part of in above embodiment is realized by way of computer program
When, which also can store in storage mediums such as server, another computer, disk, CD, flash disk or mobile hard disks
In, through downloading or copying and saving into the memory of local device, or version updating is carried out to the system of local device, when logical
When crossing the program in processor execution memory, all or part of function in above embodiment can be realized.
Use above specific case is illustrated the present invention, is merely used to help understand the present invention, not to limit
The system present invention.For those skilled in the art, according to the thought of the present invention, can also make several simple
It deduces, deform or replaces.
Claims (10)
1. a kind of detection method of vehicle steering angle, which comprises the steps of:
Image acquisition step obtains the image of at least one steering tire of vehicle in real time；
Edge detecting step carries out edge detection to current described image, obtains the bianry image comprising the tyre rim；
Edge line determines step, detects the edge line of the tire in the bianry image；
Angle detecting step obtains the steering angle and steering direction of current vehicle according to the edge line.
2. the detection method of vehicle steering angle as described in claim 1, which is characterized in that described to detect the binary map
As in tire edge line the step of include:
In the bianry image of present image, the parameter of the edge line detected according to previous frame image obtains area-of-interest；
The area of the area-of-interest is less than the area of bianry image；
The area-of-interest is detected to obtain the edge line of tire.
3. the detection method of vehicle steering angle as described in claim 1, which is characterized in that described to detect the binary map
As in tire edge line the step of include:
Hough transform is carried out to each characteristic point at edge in the bianry image；Even angle needed for Hough transform point
At n equal portions；
One group of corresponding default Hough transform engine is all made of to the angle of each equal portions while carrying out Hough transform, with detection
The edge line of the tire out.
4. the detection method of vehicle steering angle as described in claim 1, which is characterized in that the real-time acquisition vehicle is extremely
The step of image of a few steering tire includes: the image sensing being mounted on the left side mirror of vehicle or on right side mirror
Device acquires the image of corresponding steering tire in real time.
5. a kind of detection system of vehicle steering angle characterized by comprising
Image capture module, the image of at least one steering tire for obtaining vehicle in real time；
Edge detection unit obtains the two-value comprising the tyre rim for carrying out edge detection to current described image
Image；
Edge line determination unit, for detecting the edge line of the tire in the bianry image；
Angle detection unit, for obtaining the steering angle and steering direction of current vehicle according to the edge line.
6. the detection system of vehicle steering angle as claimed in claim 5, which is characterized in that the edge line determination unit tool
Body is used for:
In the bianry image of present image, the parameter of the edge line detected according to previous frame image obtains area-of-interest；
The area of the area-of-interest is less than the area of bianry image；
The area-of-interest is detected to obtain the edge line of tire.
7. the detection system of vehicle steering angle as claimed in claim 5, which is characterized in that the edge line determination unit tool
Body is used for:
Hough transform is carried out to each characteristic point at edge in the bianry image；Even angle needed for Hough transform point
At n equal portions；
One group of corresponding default Hough transform engine is all made of to the angle of each equal portions while carrying out Hough transform, with detection
The edge line of the tire out.
8. the detection system of vehicle steering angle as claimed in claim 5, which is characterized in that further include the left side for being mounted on vehicle
Imaging sensor on side mirror or on right side mirror；
Described image sensor, for acquiring the image of corresponding steering tire in real time.
9. a kind of vehicle characterized by comprising
Memory, for storing program；
Processor, for the program by executing the memory storage to realize the method as described in claim 1-4.
10. a kind of computer readable storage medium, which is characterized in that including program, described program can be executed by processor with
Realize such as method of any of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810867378.XA CN109229206A (en) | 2018-08-02 | 2018-08-02 | The detection method and system of a kind of vehicle and its steering angle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810867378.XA CN109229206A (en) | 2018-08-02 | 2018-08-02 | The detection method and system of a kind of vehicle and its steering angle |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109229206A true CN109229206A (en) | 2019-01-18 |
Family
ID=65072673
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810867378.XA Pending CN109229206A (en) | 2018-08-02 | 2018-08-02 | The detection method and system of a kind of vehicle and its steering angle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109229206A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111207699A (en) * | 2019-12-31 | 2020-05-29 | 的卢技术有限公司 | Method and system for dynamically measuring tire corner in visual mode |
| CN114771657A (en) * | 2022-06-02 | 2022-07-22 | 安徽安凯汽车股份有限公司 | Steering wheel angle detection system |
| CN115092254A (en) * | 2022-07-18 | 2022-09-23 | 苏州智加科技有限公司 | Method, device and equipment for acquiring steering angle of wheel and storage medium |
| CN116092039A (en) * | 2023-04-10 | 2023-05-09 | 智道网联科技(北京)有限公司 | Display control method and device of automatic driving simulation system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6130706A (en) * | 1998-03-25 | 2000-10-10 | Lucent Technologies Inc. | Process for determining vehicle dynamics |
| CN102620686A (en) * | 2012-02-27 | 2012-08-01 | 西北工业大学 | Two-step image detection method of vehicle steering angle movement |
| CN105678768A (en) * | 2016-01-08 | 2016-06-15 | 杭州电子科技大学 | Machine vision-based tire tread detection method |
| CN106515852A (en) * | 2016-12-19 | 2017-03-22 | 吉林大学 | Full-drive-by-wire four-wheel steering angle measuring system suitable for operating lever control |
-
2018
- 2018-08-02 CN CN201810867378.XA patent/CN109229206A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6130706A (en) * | 1998-03-25 | 2000-10-10 | Lucent Technologies Inc. | Process for determining vehicle dynamics |
| CN102620686A (en) * | 2012-02-27 | 2012-08-01 | 西北工业大学 | Two-step image detection method of vehicle steering angle movement |
| CN105678768A (en) * | 2016-01-08 | 2016-06-15 | 杭州电子科技大学 | Machine vision-based tire tread detection method |
| CN106515852A (en) * | 2016-12-19 | 2017-03-22 | 吉林大学 | Full-drive-by-wire four-wheel steering angle measuring system suitable for operating lever control |
Non-Patent Citations (4)
| Title |
|---|
| 商尔科等: "基于FPGA的实时Hough变换", 《计算机工程》 * |
| 杨隽楠等: "车辆行驶转向角的图像检测方法 ", 《西安交通大学学报》 * |
| 金浙等: "基于机器视觉的直行车道路偏离预警系统设计", 《汽车工程师》 * |
| 龚文凌等: "视频车辆自动监控系统中运动车辆捕捉算法的研究 ", 《微型电脑应用》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111207699A (en) * | 2019-12-31 | 2020-05-29 | 的卢技术有限公司 | Method and system for dynamically measuring tire corner in visual mode |
| CN114771657A (en) * | 2022-06-02 | 2022-07-22 | 安徽安凯汽车股份有限公司 | Steering wheel angle detection system |
| CN115092254A (en) * | 2022-07-18 | 2022-09-23 | 苏州智加科技有限公司 | Method, device and equipment for acquiring steering angle of wheel and storage medium |
| CN115092254B (en) * | 2022-07-18 | 2023-08-18 | 苏州智加科技有限公司 | Method, device, equipment and storage medium for acquiring steering angle of wheel |
| CN116092039A (en) * | 2023-04-10 | 2023-05-09 | 智道网联科技(北京)有限公司 | Display control method and device of automatic driving simulation system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11466988B2 (en) | Method and device for extracting key frames in simultaneous localization and mapping and smart device | |
| CN109229206A (en) | The detection method and system of a kind of vehicle and its steering angle | |
| US11657604B2 (en) | Systems and methods for estimating future paths | |
| CN112292711B (en) | Associating LIDAR data and image data | |
| CN107577988B (en) | Method, device, storage medium and program product for realizing side vehicle positioning | |
| CN102915532B (en) | Determine the method and vehicle vision system of vehicle vision system external parameter | |
| CN104590130A (en) | Rearview mirror self-adaptive adjustment method based on image identification | |
| US11948315B2 (en) | Image composition in multiview automotive and robotics systems | |
| US11436815B2 (en) | Method for limiting object detection area in a mobile system equipped with a rotation sensor or a position sensor with an image sensor, and apparatus for performing the same | |
| US10866427B2 (en) | Method and apparatus for outputting pose information | |
| US20200125861A1 (en) | Road line detection device and road line detection method | |
| CN112819864B (en) | Driving state detection method and device and storage medium | |
| CN105205459A (en) | Method and device for identifying type of image feature point | |
| US11482007B2 (en) | Event-based vehicle pose estimation using monochromatic imaging | |
| CN109740521B (en) | Parking space position determining method and device for automatic parking, electronic device and computer readable medium | |
| US10936851B2 (en) | Method and apparatus for recognizing object | |
| CN111273701B (en) | Cloud deck vision control system and control method | |
| CN103381825A (en) | Full-speed lane sensing using a plurality of cameras | |
| CN106408593A (en) | Video-based vehicle tracking method and device | |
| CN111210411B (en) | Method for detecting vanishing points in image, method for training detection model and electronic equipment | |
| Zhang et al. | Detection algorithm of takeover behavior of automatic vehicles’ drivers based on deep learning | |
| CN114037977B (en) | Road vanishing point detection method, device, equipment and storage medium | |
| CN109624855A (en) | The stop position of automatic parking determines method and device, electronic equipment and computer-readable medium | |
| CN114509089A (en) | Non-contact rail transit train speed direction mileage detection method and system | |
| WO2022186814A1 (en) | Vehicle environment modeling with a camera |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190118 |
|
| RJ01 | Rejection of invention patent application after publication |