CN104512329B - Move backward safe householder methods, device and automobile - Google Patents
Move backward safe householder methods, device and automobile Download PDFInfo
- Publication number
- CN104512329B CN104512329B CN201310451913.0A CN201310451913A CN104512329B CN 104512329 B CN104512329 B CN 104512329B CN 201310451913 A CN201310451913 A CN 201310451913A CN 104512329 B CN104512329 B CN 104512329B
- Authority
- CN
- China
- Prior art keywords
- moving objects
- image
- module
- recorded
- moving
- 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.)
- Active
Links
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60R—VEHICLES, VEHICLE FITTINGS, OR VEHICLE PARTS, NOT OTHERWISE PROVIDED FOR
- B60R1/00—Optical viewing arrangements; Real-time viewing arrangements for drivers or passengers using optical image capturing systems, e.g. cameras or video systems specially adapted for use in or on vehicles
-
- 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
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60R—VEHICLES, VEHICLE FITTINGS, OR VEHICLE PARTS, NOT OTHERWISE PROVIDED FOR
- B60R2300/00—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle
- B60R2300/80—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle characterised by the intended use of the viewing arrangement
- B60R2300/806—Details of viewing arrangements using cameras and displays, specially adapted for use in a vehicle characterised by the intended use of the viewing arrangement for aiding parking
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Mechanical Engineering (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of safe householder methods, device and automobile of moveing backward.Wherein, safe householder methods of moveing backward are comprised the following steps:The real-time video moveed backward on route by the photographic head collection vehicle for being placed in the tailstock;Moving Objects are detected in real-time video, geomery and the current location of Moving Objects is obtained;Moving Objects to detecting are tracked, and obtain the direction of motion and movement rate of Moving Objects;Geomery, current location, the direction of motion and movement rate according to Moving Objects, carries out early warning logical judgment;And according to the result of early warning logical judgment, reported to the police or do not reported to the police.The present invention can be on quick detection reversing route emergent Moving Objects and in time early warning, improve the safety in reversing process.
Description
Technical field
The invention belongs to vehicle electric field and visual processes field, and in particular to a kind of safe householder methods of reversing, fall
Car safety aid and automobile.
Background technology
With being on the increase for automobile, the accident such as touch because of the wiping that reversing is produced and be also on the increase, particularly stop in crowded
At parking lot and crossroad, as rearview mirror has very big blind area, for the Moving Objects for abruptly entering reversing route are difficult to do
Go out reaction in time, easily cause collision accident.Therefore, the Moving Objects on reversing route are detected in time, are given warning in advance, can be with
Take measures to avoid accident to user time enough.
Reversing safety assisting system mainly has two kinds at present:Radar for backing car aid system and reverse image accessory system, it
Can improve the safety of reversing to a certain extent, but still have significant limitation.The maximum spy of radar for backing car aid system
Find range from only 2m or so, this detection range cannot make timely early warning for quick close Moving Objects;In addition, reversing thunder
Up to the barrier that accurately cannot be detected less than probe installation location;And radar for backing car investigative range is limited, it is impossible to after detecting
The barrier of visor part blind area.Reverse image be only driver provide image reference, no warning function, and user without
Method pays close attention to reverse image always.
The content of the invention
It is contemplated that at least solving one of above-mentioned technical problem to a certain extent or providing at a kind of useful business
Industry is selected.For this purpose, first purpose of the present invention is to propose that a kind of monitoring scope is comprehensive, disturbance rejection is strong, prediction accuracy
The safe householder methods of high reversing.
The safe householder methods of reversing according to embodiments of the present invention, comprise the following steps:A. by being placed in the shooting of the tailstock
Real-time video on head collection vehicle reversing route;B. Moving Objects are detected in the real-time video, obtains the motion
The geomery of object and current location;C. the Moving Objects to detecting are tracked, and obtain the Moving Objects
The direction of motion and movement rate;D. the geomery according to the Moving Objects, the current location, the direction of motion
With the movement rate, early warning logical judgment is carried out;And E. is according to the result of early warning logical judgment, is reported to the police or do not reported
It is alert.
The safe householder methods of reversing according to embodiments of the present invention, at least have the advantage that:
(1) monitored using photographic head, the object in the range of detectable 10m, and the maximum detectable range of radar for backing car
Only 2m or so, the present invention have the wider more fully advantage of monitoring scope.
(2) detected with reference to inter-frame difference and background differential pair Moving Objects, illumination variation, target can be avoided cloudy
The interference of the extraneous factors such as shadow, turbulent noise, can adapt to various scene changes, and disturbance rejection is strong, prediction accuracy is high.
(3) detect the current location of the Moving Objects using distribution projection algorithm, the mode of projection is by two-dimensional space
Image is transformed into the one-dimensional space, and amount of calculation is greatly reduced, and data-handling efficiency is high, it is adaptable to real-time monitoring.
(4) consider the geomery of Moving Objects, current location, movement rate, the characteristic information of the direction of motion, sentence
It is disconnected to whether there is traffic safety hidden danger, and determine potential safety hazard grade, with true comprehensive, intelligentized advantage.
In addition, safe householder methods of moveing backward according to the above embodiment of the present invention can also have following additional technology special
Levy:
In one embodiment of the invention, also include:After step A, before step B, to the reality
When video carry out Image semantic classification.
In one embodiment of the invention, step B is specifically included:B1. judge whether present frame is the first frame;
If B2. present frame is the first frame, using the source images of present frame as the background image of present frame, into step B4;B3. such as
Fruit present frame is not the first frame, then the background image of present frame is obtained using inter-frame difference and background subtraction split-phase associated methods, is entered
Enter step B4;B4. the background image of present frame is deducted with the source images of present frame, background difference image is obtained;B5. to the back of the body
Scape difference image carries out the image binaryzation of adaptive threshold, tells moving region and non-moving areas;B6. to the motion
The background in region keeps constant, is replaced renewal to the background of the non-moving areas;And B7. adopts distribution projection algorithm
The Moving Objects in moving region are obtained, and obtains geomery and the current location of the Moving Objects.
In one embodiment of the invention, the meter of the image binaryzation process of adaptive threshold described in step B5
Calculating formula is:Wherein α is adaptation coefficient and 0 < α < 1, (i, j)
Represent the pixel of the i-th row jth row, btThe value of (i, j) for (i, j) in the background image of t present frame, bt-1(i, j) is previous
The value of (i, j), f in the background image of frametThe value of (i, j) for (i, j) in the source images of present frame, DtThe back of the body of (i, j) for t
The binaryzation result of scape difference image.
In one embodiment of the invention, step C is specifically included:C1., current kinetic object set and is provided
Record Moving Objects set, element in the current kinetic object set are to be detected by step B in present frame
The Moving Objects, the element for having recorded Moving Objects set are initialized as sky at 0 moment, by holding in following instant
Row step C3 and step C4 are updated;C2. to have recorded in Moving Objects set each element current with described one by one by described
Each element in Moving Objects set is matched, the geomery if the match is successful, to the Moving Objects
It is updated with the current location, and described has recorded in the current kinetic object set and in Moving Objects set respectively
Labelling the two elements are to match;C3. Moving Objects set has been recorded described in scanning, be not labeled if there is certain element
Match, it is believed that the Moving Objects haveing been friends in the past leave the visual field, then the element has been recorded in Moving Objects set from described
Delete described Moving Objects set has been recorded to update;C4. the current kinetic object set is scanned, if there is certain element
It is not labeled and matches, then it is assumed that there are the new Moving Objects to enter the visual field, then the element is added to and described is recorded
In Moving Objects set with update described in recorded Moving Objects set;And C5. repeat step C1 to C4, obtain each moment
It is described recorded Moving Objects set, by the Moving Objects for having been recorded in Moving Objects set it is two neighboring when
The current location carved is compared, and obtains the direction of motion and the movement rate of the Moving Objects.
In one embodiment of the invention, while performing step C, also include:Record the Moving Objects
Credibility, the credibility are used to aid in having recorded Moving Objects set described in renewal, wherein, when continuous multiple frames image detection is arrived
During the Moving Objects, the credibility of the Moving Objects gradually adds up;When continuous multiple frames image detection is less than the fortune
During dynamic object, the credibility of the Moving Objects is gradually successively decreased.
For this purpose, second object of the present invention is to propose that a kind of monitoring scope is comprehensive, disturbance rejection is strong, prediction accuracy
High reversing safety aid.
Reversing safety aid according to embodiments of the present invention, including:Video acquisition module, the video acquisition module
Including the photographic head for being placed in the tailstock, for the real-time video on collection vehicle reversing route;Moving Objects detection module, the fortune
Dynamic obj ect detection module is connected with the video acquisition module, for Moving Objects are detected in the real-time video, is obtained
The geomery of the Moving Objects and current location;Moving object tracking module, the moving object tracking module with it is described
Moving Objects detection module is connected, and is tracked for the Moving Objects to detecting, obtains the fortune of the Moving Objects
Dynamic direction and movement rate;Early warning logic judgment module, the early warning logic judgment module and the moving object tracking module
It is connected, for the geomery according to the Moving Objects, the current location, the direction of motion and motion speed
Rate, carries out early warning logical judgment;And alarm, the alarm is connected with the early warning logic judgment module, according to described
The judged result of early warning logic judgment module is reported to the police or is not reported to the police.
Reversing safety aid according to embodiments of the present invention, at least has the advantage that:
(1) video acquisition module monitored using photographic head, can detect the object in the range of 10m, and radar for backing car
Maximum detectable range is only 2m or so, and the present invention is with the wider more fully advantage of monitoring scope.
(2), in Moving Objects detection module, detected with reference to inter-frame difference and background differential pair Moving Objects, can be kept away
Exempt from the interference of the extraneous factors such as illumination variation, target shadow, turbulent noise, can adapt to various scene changes, disturbance rejection is strong,
Prediction accuracy is high.
(3), in moving object tracking module, the current location of the Moving Objects, projection are detected using distribution projection algorithm
Mode the image of two-dimensional space is transformed into into the one-dimensional space, amount of calculation is greatly reduced, and data-handling efficiency is high, it is adaptable in real time
Monitoring.
(4) consider the geomery of Moving Objects, current location, movement rate, the characteristic information of the direction of motion, sentence
It is disconnected to whether there is traffic safety hidden danger, and determine potential safety hazard grade, with true comprehensive, intelligentized advantage.
In addition, safe householder methods of moveing backward according to the above embodiment of the present invention can also have following additional technology special
Levy:
In one embodiment of the invention, also include:Image pre-processing module, described image pretreatment module are located at institute
State between video acquisition module and the Moving Objects detection module, for the real-time video and carrying out Image semantic classification.
In one embodiment of the invention, the Moving Objects detection module is specifically included:Background image acquisition module,
Wherein, if present frame is the first frame, background of the background image acquisition module using the source images of present frame as present frame
Image, if present frame is not the first frame, the background image acquisition module adopts inter-frame difference and background subtraction split-phase combination side
Method obtains the background image of present frame;Background difference image acquisition module, the background difference image acquisition module and the back of the body
Scape image collection module is connected, and the background image for the source images of present frame to be deducted present frame obtains background difference image;
Image binaryzation module, described image binarization block are connected with the background difference image acquisition module, for the back of the body
Scape difference image carries out the image binaryzation of adaptive threshold, tells moving region and non-moving areas;Context update module,
The context update module is connected with described image binarization block, keeps constant for the background to the moving region, right
The background of the non-moving areas is replaced renewal;And size positions acquisition module, size positions acquisition module with it is described
Image binaryzation module is connected, and for the Moving Objects in moving region are obtained using distribution projection algorithm, and obtains the fortune
The geomery of dynamic object and current location.
In one embodiment of the invention, the image binaryzation of adaptive threshold described in described image binarization block
The computing formula of process is:Wherein α is adaptation coefficient and 0 < α <
1, (i, j) represents the pixel of the i-th row jth row, btThe value of (i, j) for (i, j) in the background image of t present frame, bt-1(i,
J) it is the value of (i, j) in the background image of former frame, ftThe value of (i, j) for (i, j) in the source images of present frame, Dt(i, j) is t
The binaryzation result of the background difference image at moment.
In one embodiment of the invention, the moving object tracking module is specifically included:Current kinetic object set
Internal memory, for storing current kinetic object set, the element in the current kinetic object set is by described in present frame
The Moving Objects that Moving Objects detection module is detected;Moving Objects set internal memory has been recorded, fortune has been recorded for storage
Dynamic object set, the element for having recorded Moving Objects set are initialized as sky at 0 moment, in following instant constantly more
Newly;Matching module, the matching module respectively with the current kinetic object set internal memory and recorded in Moving Objects set
Deposit connected, for by it is described recorded each element in Moving Objects set one by one with the current kinetic object set in it is each
Individual element is matched, if the match is successful, the geomery and the current location of the Moving Objects is carried out
Update, and respectively in the current kinetic object set and described the two elements of labelling that recorded in Moving Objects set for
Matching;Scanning update module, the scanning update module respectively with the current kinetic object set internal memory and recorded motion
Object set internal memory is connected, for having recorded Moving Objects set and being updated to described, wherein, the scanning update module is swept
Moving Objects set is recorded described in retouching, has not been labeled if there is certain element and is matched, it is believed that the motion haveing been friends in the past is right
As leaving the visual field, then the element is deleted to update in Moving Objects set by the scanning update module from described record
State;The current kinetic object set is scanned with the scanning update module, if there is certain
Element is not labeled and matches, then it is assumed that have the new Moving Objects to enter the visual field, then the scanning update module just institute
Stating element and being added to described recorded in Moving Objects set Moving Objects set have been recorded described in updating;And kinematic parameter
Acquisition module, is connected with the Moving Objects set internal memory that recorded, for by the institute for having been recorded in Moving Objects set
The current location for stating the two neighboring moment of Moving Objects is compared, and obtains the direction of motion of the Moving Objects
With the movement rate.
In one embodiment of the invention, also include in the moving object tracking module:Credibility internal memory, it is described can
Reliability internal memory is used for the credibility for recording the Moving Objects, and the credibility is used to aid in having recorded Moving Objects described in renewal
Set, wherein, when continuous multiple frames image detection is to the Moving Objects, the credibility of the Moving Objects is gradually tired out
Plus;When continuous multiple frames image detection is less than the Moving Objects, the credibility of the Moving Objects is gradually successively decreased.
For this purpose, third object of the present invention is to propose that a kind of monitoring scope is comprehensive, disturbance rejection is strong, prediction accuracy
High automobile.
Automobile according to embodiments of the present invention, including any one reversing safety aid that the present invention is disclosed above.
As the automobile of the embodiment of the present invention is with the similar knot to the reversing safety aid of the embodiment of the present invention
Structure, therefore also have the advantages that similar.
The additional aspect and advantage of the present invention will be set forth in part in the description, and partly will become from the following description
Obtain substantially, or recognized by the practice of the present invention.
Description of the drawings
The above-mentioned and/or additional aspect and advantage of the present invention will become from the description with reference to accompanying drawings below to embodiment
It is substantially and easy to understand, wherein:
Fig. 1 is the schematic flow sheet of the safe householder methods of reversing of the embodiment of the present invention;
Fig. 2 is the Moving Objects detection process schematic flow sheet of the safe householder methods of reversing of the embodiment of the present invention;
Fig. 3 is the moving object tracking process flow schematic diagram of the safe householder methods of reversing of the embodiment of the present invention;With
Fig. 4 is the structured flowchart of the reversing safety aid of the embodiment of the present invention.
Specific embodiment
Embodiments of the invention are described below in detail, the example of the embodiment is shown in the drawings, wherein from start to finish
Same or similar label represents same or similar element or the element with same or like function.Below with reference to attached
The embodiment of figure description is exemplary, it is intended to for explaining the present invention, and be not considered as limiting the invention.
As shown in figure 1, the safe householder methods of the reversing of the embodiment of the present invention, specifically may include steps of:
A. the real-time video moveed backward on route by the photographic head collection vehicle for being placed in the tailstock.
Specifically:The situation on vehicle backing route is shot by being placed in the photographic head of tail of the body, photographic head
Effective resolution be 640 × 480, frame per second be 30 frames/second.As photographic head installation site and angle can directly affect shooting area
Domain, misoperation are then possible to produce blind area, therefore the installation process is rather important.Specific installation site and angle be not according to
With vehicle and vehicle body size and it is different, it is preferable that photographic head installation direction is installed using inclining obliquely, so can be with
Real time status on vehicle backing route are obtained preferably.
Preferably, after above-mentioned steps A are completed, need to real-time video and carry out Image semantic classification.Specifically:Collection
After complete real-time video, the analogue signal that photographic head is input into need to be decoded as by Video Decoder the digital video of YUV, A/ is completed
D changes and is input into dsp chip, and image is preserved by DSP by PPI interfaces.After DSP is gathered and preserved image, using 3 × 3 height
This wave filter carries out image denoising and smooth pretreatment, improves picture quality.
B. Moving Objects are detected in real-time video, obtains geomery and the current location of Moving Objects.
Moving Objects detection technique is exactly that separation is on a frame-by-frame basis detected in the region of motion change in sequence of video images
Out, the impact of background image is removed so as to find target prospect part interested.In actual environment, due to illumination variation,
The interference of the extraneous factors such as target shadow, turbulent noise causes the accurate detection of Moving Objects to become extremely difficult.The present embodiment
The detection process of Moving Objects combine Fig. 2, specifically may include steps of:
B1. judge whether present frame is the first frame.
If B2. present frame is the first frame, using the source images of present frame as the background image of present frame, into step
B4。
Specifically, remember the corresponding moment t=0 of the first frame, then have Back (t)=SourceImage (t), t=0.Wherein Back
T () represents t present frame(That is t+1 frames)Background image, SourceImage (t) represent t present frame source figure
Picture.
If B3. present frame is not the first frame, present frame is obtained using inter-frame difference and background subtraction split-phase associated methods
Background image, into step B4.
Traditional background difference algorithm is made in the meansigma methodss that the first frame or former frames are chosen in Moving Objects detection at the beginning
For the initial model of background image, but for Moving Objects are contained in initial frame, this is clearly inaccurate, when in subsequent treatment
Context update speed can cause to have a strong impact on to detection when very slow.Therefore the present embodiment adopts inter-frame difference and background subtraction split-phase to tie
Close, to improve accuracy, specific formula for calculation is as follows:
As can be seen from the above equation, when SourceImage (t)-Back (t-1) difference is less, Back(t)Almost with Back(t
+1)It is equal.This means that the method has the characteristic of low pass, Moving Objects are regarded as random noise, carried out background image
Obtain, algorithm complex is low.From above formula it can also be seen that when t becomes big, 1/t is gradually reduced so formula latter half can be with
Ignore, calculating speed is fast, operand reduces, so as to meet the real-time of object detection.
B4. the background image of present frame is deducted with the source images of present frame, background difference image is obtained.
Specifically, in t, in remembering the source images of present frame, (i, j) picture element point chromatic value is ft(i, j), present frame
In background image, (i, j) picture element point chromatic value is bt(i, j), then (i, j) pixel color in the background difference image of present frame
Color value is dt(i,j)=ft(i,j)-bt(i,j)。
B5. the image binaryzation of adaptive threshold is carried out to background difference image, moving region and non-athletic area is told
Domain.
Specifically, the computing formula of image binaryzation is:After binaryzation, value is 1
Region mean that the source images and background image of present frame differ greatly, belong to moving region,;After binaryzation, value is 0
Region means that the source images of present frame and background image difference are less, belongs to non-moving areas.
It should be noted that in non-moving areas, the fine difference of source images and background image is due in reversing process
Photographic head visual angle change causes, and such as one tree can also occur slow position motion in back-up video, and which belongs to non-fortune
Dynamic region.In moving region, on the one hand differing greatly in source images and background image, the difference are photographic head in reversing process
The relative motion that visual angle change causes, be on the other hand due to the absolute movement of Moving Objects itself, such as one dart it is little
There is visibility point motion in Canis familiaris L. in back-up video, which belongs to moving region.
It should be noted that T is current source picture and the binary-state threshold after current background calculus of differences.Image after difference
Noise profile is very big and the identical characteristics with Gauss distribution, and Moving Objects region does not possess this characteristic, therefore passes through
Calculate the estimated value that the average of difference image, standard deviation can obtain the average of corresponding noise profile, standard deviation.According to general
Rate Statistics, the stochastic variable of Gauss distribution meet 3 δ principles, and threshold value T tried to achieve hence with 3 δ principles is carried out at binaryzation
Reason can be realized at utmost suppressing the noise of difference image, so as to substantially reduce the interference of noise, determine moving region.
B6. constant is kept to the background of moving region, renewal is replaced to the background of non-moving areas.
The computing formula of the step can be:Wherein α is adaptive
Answer coefficient and 0 < α < 1.Other symbols hereinbefore have lexical or textual analysis, wherein (i, j) represents the pixel of the i-th row jth row, bt(i,
J) it is the value of (i, j) in the background image of t present frame, bt-1The value of (i, j) for (i, j) in the background image of former frame, ft
The value of (i, j) for (i, j) in the source images of present frame, DtThe binaryzation result of (i, j) for the background difference image of t.
Step B6 is substantially and further context update is done on the basis of step B3 is to context update.If only according to
By the background update method of step B3 description, with the increase of t, background model tends towards stability.And in practical application, with vehicle
The movement background moment changing, need to be updated non-moving areas background.The method described by step B6, can
To update constantly the background of non-moving areas, and the detection of Moving Objects is not affected.
B7. the Moving Objects in moving region are obtained using distribution projection algorithm, and obtains the geomery of Moving Objects
And current location.
Substep projection algorithm refers to that do vertical projection to image first, each Moving Objects can be produced in projection result
Corresponding crest, therefore the number of Moving Objects and each fortune in image of interest region can be detected by vertical projection
The horizontal boundary of dynamic object.Then floor projection is done respectively in its horizontal boundary to each Moving Objects, its projection result can
With the vertical border of each Moving Objects of accurate response, so as to detect the boundary position of each Moving Objects, it is determined that motion is right
The geomery of elephant and current location.The image of two-dimensional space is transformed into the one-dimensional space by the mode of distribution projection, and data volume is big
It is big to reduce, it is adaptable to the real-time detection of embedded system.
C. the Moving Objects to detecting are tracked, and obtain the direction of motion and movement rate of Moving Objects.
Moving Objects are tracked and are referred to:Create based on position, size, speed, direction etc. in continuous image interframe
Corresponding matching and record Moving Objects movement locus in the video sequence about feature.As shown in figure 3, step C is concrete
May include steps of:
C1., current kinetic object set is provided and Moving Objects set has been recorded, the element in current kinetic object set
For the Moving Objects detected by step B in present frame, the element for having recorded Moving Objects set was initialized as at 0 moment
Sky, is updated by execution step C3 and step C4 in following instant.
C2. in having been recorded Moving Objects set, each element is entered with each element in current kinetic object set one by one
Row matching, if the match is successful, is updated to the geomery and current location of Moving Objects, and respectively in current kinetic
Object set and in having recorded Moving Objects set the two elements of labelling to match.
C3. scan and recorded Moving Objects set, be not labeled if there is certain element and match, it is believed that the fortune haveing been friends in the past
Dynamic object leaves the visual field, then by element from record delete in Moving Objects set with update recorded Moving Objects set.
C4. current kinetic object set is scanned, is not labeled if there is certain element and is matched, then it is assumed that have new fortune
Dynamic object enters the visual field, then element is added to record in Moving Objects set with update recorded Moving Objects set.
It should be noted that step C3 and step C4 can with successively, it is rear first or while order carry out.
C5. repeat C1 to C4, obtain the Moving Objects set of record at each moment, recorded in Moving Objects set
The current location at two neighboring moment of Moving Objects be compared, obtain the direction of motion and movement rate of Moving Objects.
Above-mentioned moving object tracking method it be feature based moving object tracking method, accuracy is higher, with one
Fixed noise immunity.
An example of the invention, while execution step C, also includes:The credibility of record Moving Objects, it is credible
Spend Moving Objects set has been recorded for aiding in updating.Wherein, when continuous multiple frames image detection is to Moving Objects, Moving Objects
Credibility gradually add up;When continuous multiple frames image detection is less than Moving Objects, the credibility of Moving Objects is gradually successively decreased.When
After Moving Objects have credibility parameter, the credibility is just may be referred in execution step C3.For example, scanning has recorded fortune
Dynamic object set, certain element of discovery presence first is labeled to be matched(Detect and recorded in Moving Objects set
The unmarked matching first of a certain element), it is meant that tracked with losing the corresponding Moving Objects of the element first.Now practical situation
It is not necessarily that the Moving Objects are thoroughly disappeared in the visual field, it is also possible to be error.Therefore the element should not be deleted at once, but subtract
The credibility of the little element, until credibility be less than believability threshold when, just think that Moving Objects really leave, just by element from
Delete in having recorded Moving Objects set.By means of the auxiliary of credibility so that the tracking result of the present embodiment is truer more accurate
Really.
D. geomery according to Moving Objects, current location, the direction of motion and movement rate, carry out early warning logic and sentence
It is disconnected.
First, the geomery of the Moving Objects for being obtained according to above-mentioned steps B and step C, current location, the direction of motion and
The information such as movement rate, the movement locus of the following Moving Objects in a short time of prediction.Then, comprehensive descision is with the presence or absence of driving peace
Full hidden danger, and determine potential safety hazard grade.For example, it is predicted that the relative velocity that Moving Objects volume is larger for certain, which is with vehicle
It is larger, and its relative distance with vehicle is less than secure threshold, then mean to be likely to acutely collision in the future, now
Judge this situation for severe compromise.Again for example, it is predicted that certain Moving Objects small volume, its relative velocity with vehicle
It is less, and which is slightly larger than secure threshold with the relative distance of vehicle, then mean to be possible to slight collision in the future, this
When judge that this situation is common potential safety hazard.Those skilled in the art can neatly realize step D according to practical situation.
E. reported to the police or do not reported to the police according to the result of early warning logical judgment.
Report to the police and can adopt the combination of both audible alarm, photoelectric alarm, audible alarm and photoelectric alarm, or other shapes
Formula, warns to user.
The safe householder methods of reversing according to embodiments of the present invention, at least have the advantage that:
(1) monitored using photographic head, the object in the range of detectable 10m, and the maximum detectable range of radar for backing car
Only 2m or so, the present invention have the wider more fully advantage of monitoring scope.
(2) detected with reference to inter-frame difference and background differential pair Moving Objects, illumination variation, target can be avoided cloudy
The interference of the extraneous factors such as shadow, turbulent noise, can adapt to various scene changes, and disturbance rejection is strong, prediction accuracy is high.
(3) current location and position of Moving Objects are detected using distribution projection algorithm, the mode of projection is by two-dimensional space
Image be transformed into the one-dimensional space, amount of calculation is greatly reduced, and data-handling efficiency is high, it is adaptable to real-time monitoring.
(4) created based on position, size, speed, direction etc. in continuous image interframe by moving object tracking algorithm
Corresponding matching and record Moving Objects movement locus in the video sequence about feature, accuracy are high, with certain
Noise immunity.
(5) consider the geomery of Moving Objects, current location, movement rate, the characteristic information of the direction of motion, sentence
It is disconnected to whether there is traffic safety hidden danger, and determine potential safety hazard grade, with true comprehensive, intelligentized advantage.
The present invention proposes a kind of reversing safety aid, as shown in figure 4, can include:Video acquisition module 100, fortune
Dynamic obj ect detection module 200, moving object tracking module 300 and early warning logic judgment module 400.Video acquisition module 100 is wrapped
The photographic head for being placed in the tailstock is included, for obtaining the real-time video on vehicle backing route.Moving Objects detection module 200 and video
Acquisition module 100 is connected, and for Moving Objects are detected in real-time video, obtains the geomery and present bit of Moving Objects
Put.Moving object tracking module 300 is connected with Moving Objects detection module 200, for the Moving Objects to detecting carry out with
Track, obtains the direction of motion and movement rate of Moving Objects.Early warning logic judgment module 400 and moving object tracking module 300
It is connected, for the geomery according to Moving Objects, current location, the direction of motion and movement rate, carries out early warning logical judgment.
Alarm 500 is connected with early warning logic judgment module 400, is reported to the police according to the judged result of early warning logic judgment module 400
Or do not report to the police.
An example of the invention, reversing safety aid can also include:Image pre-processing module 600.Should
Image pre-processing module 600 is located between video acquisition module 100 and Moving Objects detection module 200, for collecting
Real-time video simultaneously carries out Image semantic classification.
An example of the invention, Moving Objects detection module 200 specifically can include background image acquisition module,
Background difference image acquisition module, image binaryzation module, context update module and size positions acquisition module.Background image is obtained
In delivery block, if present frame is the first frame, background of the background image acquisition module using the source images of present frame as present frame
Image, if present frame is not the first frame, background image acquisition module is obtained using inter-frame difference and background subtraction split-phase associated methods
Obtain the background image of present frame.Background difference image acquisition module is connected with background image acquisition module, for by present frame
Source images deduct the background image of present frame and obtain background difference image.Image binaryzation module obtains mould with background difference image
Block is connected, and for the image binaryzation of adaptive threshold is carried out to background difference image, tells moving region and non-athletic area
Domain.Context update module is connected with image binaryzation module, keeps constant for the background to moving region, to non-moving areas
Background be replaced renewal.Size positions acquisition module is connected with image binaryzation module, for using distribution projection algorithm
The Moving Objects in moving region are obtained, and obtains geomery and the current location of Moving Objects.
An example of the invention, the calculating of the image binaryzation process of adaptive threshold in image binaryzation module
Formula can be:Wherein α is adaptation coefficient and 0 < α < 1, (i,
J) pixel of the i-th row jth row, b are representedtThe value of (i, j) for (i, j) in the background image of t present frame, bt-1(i, j) is front
The value of (i, j), f in the background image of one frametThe value of (i, j) for (i, j) in the source images of present frame, Dt(i, j) is t
The binaryzation result of background difference image.
An example of the invention, moving object tracking module 300 can specifically include:Current kinetic object set
Internal memory, Moving Objects set internal memory, matching module, scanning update module and kinematic parameter acquisition module are recorded.Current kinetic
Object set internal memory is used to store current kinetic object set.Element in current kinetic object set be present frame in by fortune
The Moving Objects that dynamic obj ect detection module is detected.Recorded Moving Objects set internal memory for storage recorded Moving Objects collection
Close.The element for having recorded Moving Objects set is initialized as sky at 0 moment, constantly updates in following instant.Matching module point
Not with current kinetic object set internal memory and recorded Moving Objects set internal memory and be connected, for being recorded Moving Objects set
In each element matched with each element in current kinetic object set one by one.It is if the match is successful, right to moving
The geomery of elephant and current location are updated, and have recorded in current kinetic object set and in Moving Objects set respectively
Labelling the two elements are to match.Scanning update module respectively with current kinetic object set internal memory and recorded Moving Objects
Set internal memory is connected, for being updated to recording Moving Objects set.The renewal process in two kinds of situation, the first feelings
Condition, scanning update module scanning have recorded Moving Objects set, are not labeled if there is certain element and match, it is believed that have been friends in the past
Moving Objects leave the visual field, then scan update module by element from record delete in Moving Objects set with update record
Moving Objects set.Second situation, scanning update module scanning current kinetic object set, if there is certain element not by
Labelling is matched, then it is assumed that have new Moving Objects to enter the visual field, then scanning update module just element is added to
Moving Objects set has been recorded to update in dynamic object set.Kinematic parameter acquisition module with record Moving Objects set internal memory
It is connected, the current location for being recorded the two neighboring moment of the Moving Objects in Moving Objects set is compared, and obtains
Obtain the direction of motion and movement rate of Moving Objects.
An example of the invention, can also include in moving object tracking module 300:Credibility internal memory.This can
Reliability internal memory is used for the credibility for recording Moving Objects, and credibility is used to aid in updating to have recorded Moving Objects set.When continuous
When multiple image detects Moving Objects, the credibility of Moving Objects gradually adds up;When continuous multiple frames image detection is less than motion
During object, the credibility of Moving Objects is gradually successively decreased.When scanning update module carries out the first to recording Moving Objects set
During the renewal of situation, credibility is may be referred to.For example, scanning update module is detected and records a certain in Moving Objects set
Element unmarked matching first, that is, mean to be tracked with losing the corresponding Moving Objects of the element first.Now practical situation is not
Must be that the Moving Objects are thoroughly disappeared in the visual field, it is also possible to be error.Therefore the element should not be deleted at once, but is reduced
The credibility of the element, until when credibility is less than believability threshold, just thinking that Moving Objects really leave, just by element from
Delete in record Moving Objects set.By means of the auxiliary of credibility so that the tracking result of the present invention is truer more accurate.
The reversing safety aid of the embodiment of the present invention, for the fortune on reversing route is suddenly appeared in reversing process
Dynamic object detection, give warning in advance guarantee safety.
Reversing safety aid according to embodiments of the present invention, at least has the advantage that:
(1) monitored using photographic head, the object in the range of detectable 10m, and the maximum detectable range of radar for backing car
Only 2m or so, the present invention have the wider more fully advantage of monitoring scope.
(2) detected with reference to inter-frame difference and background differential pair Moving Objects, illumination variation, target can be avoided cloudy
The interference of the extraneous factors such as shadow, turbulent noise, can adapt to various scene changes, and disturbance rejection is strong, prediction accuracy is high.
(3) current location and position of Moving Objects are detected using distribution projection algorithm, the mode of projection is by two-dimensional space
Image be transformed into the one-dimensional space, amount of calculation is greatly reduced, and data-handling efficiency is high, it is adaptable to real-time monitoring.
(4) created based on position, size, speed, direction etc. in continuous image interframe by moving object tracking algorithm
Corresponding matching and record Moving Objects movement locus in the video sequence about feature, accuracy are high, with certain
Noise immunity.
(5) consider the geomery of Moving Objects, current location, movement rate, the characteristic information of the direction of motion, sentence
It is disconnected to whether there is traffic safety hidden danger, and determine potential safety hazard grade, with true comprehensive, intelligentized advantage.
The present invention also propose a kind of monitoring scope comprehensively, the automobile that disturbance rejection is strong, prediction accuracy is high.According to the present invention
The automobile of embodiment, including any one reversing safety aid that the present invention is disclosed above.
As the automobile of the embodiment of the present invention is with the similar knot to the reversing safety aid of the embodiment of the present invention
Structure, therefore also have the advantages that similar.
In flow chart or here any process described otherwise above or method description are construed as, expression includes
It is one or more for realizing specific logical function or process the step of the module of code of executable instruction, fragment or portion
Point, and the scope of the preferred embodiment of the present invention includes other realization, wherein the suitable of shown or discussion can not be pressed
Sequence, including according to involved function by it is basic simultaneously in the way of or in the opposite order, carry out perform function, this should be of the invention
Embodiment person of ordinary skill in the field understood.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
Example ", or the description of " some examples " etc. mean specific features with reference to the embodiment or example description, structure, material or spy
Point is contained at least one embodiment or example of the present invention.In this manual, to the schematic representation of above-mentioned term not
Identical embodiment or example are referred to necessarily.And, the specific features of description, structure, material or feature can be any
One or more embodiments or example in combine in an appropriate manner.
Although embodiments of the invention have been shown and described above, it is to be understood that above-described embodiment is example
Property, it is impossible to limitation of the present invention is interpreted as, one of ordinary skill in the art is in the principle and objective without departing from the present invention
In the case of above-described embodiment can be changed within the scope of the invention, change, replace and modification.
Claims (11)
1. safe householder methods of a kind of reversing, it is characterised in that comprise the following steps:
A. the real-time video moveed backward on route by the photographic head collection vehicle for being placed in the tailstock;
B. Moving Objects are detected in the real-time video with reference to inter-frame difference and background difference, obtains the Moving Objects
Geomery and current location, wherein, step B is specifically included:B1. judge whether present frame is the first frame;If B2. worked as
Previous frame is the first frame, then using the source images of present frame as the background image of present frame, into step B4;If B3. present frame
It is not the first frame, then the background image of present frame is obtained using inter-frame difference and background subtraction split-phase associated methods, into step B4;
B4. the background image of present frame is deducted with the source images of present frame, background difference image is obtained;B5. to the background difference diagram
Image binaryzation as carrying out adaptive threshold, tells moving region and non-moving areas;B6. the back of the body to the moving region
Scape keeps constant, is replaced renewal to the background of the non-moving areas;B7. moving region is obtained using distribution projection algorithm
Interior Moving Objects, and obtain geomery and the current location of the Moving Objects;
C. the Moving Objects to detecting are tracked, and obtain the direction of motion and movement rate of the Moving Objects;
D. the geomery according to the Moving Objects, the current location, the direction of motion and the movement rate,
Carry out early warning logical judgment;And
E. the result according to the early warning logical judgment, is reported to the police or is not reported to the police.
2. safe householder methods of reversing according to claim 1, it is characterised in that also include:After step A,
Before step B, to the real-time video and Image semantic classification is carried out.
3. safe householder methods of reversing according to claim 1, it is characterised in that adaptive thresholding described in step B5
The computing formula of the image binaryzation process of value is:Wherein α is certainly
Accommodation coefficient and 0 < α < 1, (i, j) represent the pixel of the i-th row jth row, btIn the background image of (i, j) for t present frame
The value of (i, j), bt-1The value of (i, j) for (i, j) in the background image of former frame, ft(i, j) in the source images of present frame (i,
J) value, DtThe binaryzation result of (i, j) for the background difference image of t.
4. safe householder methods of reversing according to claim 1, it is characterised in that step C is specifically included:
C1., current kinetic object set is provided and Moving Objects set has been recorded, the element in the current kinetic object set
For the Moving Objects detected by step B in present frame, the element for having recorded Moving Objects set is at 0
Quarter is initialized as sky, is updated by execution step C3 and step C4 in following instant;
C2. by it is described recorded each element in Moving Objects set one by one with the current kinetic object set in each yuan
Element is matched, if the match is successful, the geomery and the current location of the Moving Objects is updated,
And respectively in the current kinetic object set and described the two elements of labelling that recorded in Moving Objects set to match;
C3. Moving Objects set has been recorded described in scanning, be not labeled if there is certain element and match, it is believed that the institute haveing been friends in the past
State Moving Objects and leave the visual field, then by the element from it is described recorded to delete in Moving Objects set described recorded with updating
Moving Objects set;
C4. the current kinetic object set is scanned, is not labeled if there is certain element and is matched, then it is assumed that have new institute
State Moving Objects and enter the visual field, then the element is added to and described has been recorded in Moving Objects set to be recorded described in updating
Moving Objects set;And
C5. repeat step C1 to C4, obtains the described of each moment and has recorded Moving Objects set, by it is described record motion it is right
The current location at the two neighboring moment of the Moving Objects in as set is compared, and obtains the Moving Objects
The direction of motion and the movement rate.
5. safe householder methods of reversing according to claim 4, it is characterised in that while performing step C, also wrap
Include:The credibility of the Moving Objects is recorded, and the credibility is used to aid in Moving Objects set having been recorded described in renewal, its
In,
When continuous multiple frames image detection is to the Moving Objects, the credibility of the Moving Objects gradually adds up;
When continuous multiple frames image detection is less than the Moving Objects, the credibility of the Moving Objects is gradually successively decreased.
6. a kind of reversing safety aid, it is characterised in that include:
Video acquisition module, the video acquisition module include the photographic head for being placed in the tailstock, and the video acquisition module is used to adopt
Real-time video on collection vehicle backing route;
Moving Objects detection module, the Moving Objects detection module are connected with the video acquisition module, for combining interframe
Difference and background difference detect Moving Objects in the real-time video, obtain the geomery of the Moving Objects and current
Position, wherein, the Moving Objects detection module is specifically included:Background image acquisition module, wherein, if present frame is first
Frame, background image of the background image acquisition module using the source images of present frame as present frame, if present frame is not
One frame, the background image acquisition module adopt inter-frame difference and background subtraction split-phase associated methods to obtain the Background of present frame
Picture;Background difference image acquisition module, the background difference image acquisition module are connected with the background image acquisition module, are used
Background difference image is obtained in the background image that the source images of present frame are deducted present frame;Image binaryzation module, the figure
As binarization block is connected with the background difference image acquisition module, for carrying out adaptive thresholding to the background difference image
The image binaryzation of value, tells moving region and non-moving areas;Context update module, the context update module with it is described
Image binaryzation module is connected, and keeps constant for the background to the moving region, the background of the non-moving areas is entered
Row is replaced and is updated;Size positions acquisition module, size positions acquisition module are connected with described image binarization block, for adopting
Distribution projection algorithm obtains the Moving Objects in moving region, and obtains geomery and the current location of the Moving Objects;
Moving object tracking module, the moving object tracking module are connected with the Moving Objects detection module, for inspection
The Moving Objects measured are tracked, and obtain the direction of motion and movement rate of the Moving Objects;
Early warning logic judgment module, the early warning logic judgment module are connected with the moving object tracking module, for basis
The geomery of the Moving Objects, the current location, the direction of motion and the movement rate, carry out early warning and patrol
Collect and judge;And
Alarm, the alarm are connected with the early warning logic judgment module, according to sentencing for the early warning logic judgment module
Disconnected result is reported to the police or is not reported to the police.
7. reversing safety aid according to claim 6, it is characterised in that also include:Image pre-processing module, institute
State image pre-processing module to be located between the video acquisition module and the Moving Objects detection module, for described real-time
Video simultaneously carries out Image semantic classification.
8. reversing safety aid according to claim 6, it is characterised in that described in described image binarization block
The computing formula of the image binaryzation process of adaptive threshold is:Its
Middle α is adaptation coefficient and 0 < α < 1, and (i, j) represents the pixel of the i-th row jth row, btThe background of (i, j) for t present frame
The value of (i, j), b in imaget-1The value of (i, j) for (i, j) in the background image of former frame, ftThe source images of (i, j) for present frame
In (i, j) value, DtThe binaryzation result of (i, j) for the background difference image of t.
9. reversing safety aid according to claim 6, it is characterised in that the moving object tracking module is concrete
Including:
Current kinetic object set internal memory, for storing current kinetic object set, the unit in the current kinetic object set
The plain Moving Objects to be detected by the Moving Objects detection module in present frame;
Moving Objects set internal memory has been recorded, Moving Objects set has been recorded for storage, it is described to have recorded Moving Objects set
Element be initialized as sky at 0 moment, in following instant constantly update;
Matching module, the matching module respectively with the current kinetic object set internal memory and recorded in Moving Objects set
Deposit connected, for by it is described recorded each element in Moving Objects set one by one with the current kinetic object set in it is each
Individual element is matched, if the match is successful, the geomery and the current location of the Moving Objects is carried out
Update, and respectively in the current kinetic object set and described the two elements of labelling that recorded in Moving Objects set for
Matching;
Scanning update module, the scanning update module respectively with the current kinetic object set internal memory and recorded move it is right
It is connected as gathering internal memory, for having recorded Moving Objects set and being updated to described, wherein,
The scanning update module scanning is described to have recorded Moving Objects set, labeled if there is certain element
Match somebody with somebody, it is believed that the Moving Objects haveing been friends in the past leave the visual field, then the element has been recorded fortune from described by the scanning update module
Delete in dynamic object set described Moving Objects set has been recorded to update;With
The scanning update module scans the current kinetic object set, is not labeled if there is certain element and matches,
Then think there are the new Moving Objects to enter the visual field, then the scanning just described element of update module is added to and described remembers
Record Moving Objects set in update described in recorded Moving Objects set;And
Kinematic parameter acquisition module, is connected with the Moving Objects set internal memory that recorded, for by it is described record motion it is right
The current location at the two neighboring moment of the Moving Objects in as set is compared, and obtains the Moving Objects
The direction of motion and the movement rate.
10. reversing safety aid according to claim 9, it is characterised in that in the moving object tracking module
Also include:Credibility internal memory, the credibility internal memory are used for the credibility for recording the Moving Objects, and the credibility is used for auxiliary
Help Moving Objects set has been recorded described in renewal, wherein,
When continuous multiple frames image detection is to the Moving Objects, the credibility of the Moving Objects gradually adds up;
When continuous multiple frames image detection is less than the Moving Objects, the credibility of the Moving Objects is gradually successively decreased.
11. a kind of automobiles, it is characterised in that including the reversing safety aid any one of claim 6-10.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310451913.0A CN104512329B (en) | 2013-09-27 | 2013-09-27 | Move backward safe householder methods, device and automobile |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310451913.0A CN104512329B (en) | 2013-09-27 | 2013-09-27 | Move backward safe householder methods, device and automobile |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104512329A CN104512329A (en) | 2015-04-15 |
CN104512329B true CN104512329B (en) | 2017-04-05 |
Family
ID=52788373
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310451913.0A Active CN104512329B (en) | 2013-09-27 | 2013-09-27 | Move backward safe householder methods, device and automobile |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104512329B (en) |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106608221B (en) * | 2015-10-26 | 2019-06-25 | 比亚迪股份有限公司 | The detection system and method for vehicle blind zone |
CN105946718B (en) * | 2016-06-08 | 2019-04-05 | 深圳芯智汇科技有限公司 | The method of car-mounted terminal and its switching display reverse image |
CN107644528A (en) * | 2017-08-02 | 2018-01-30 | 浙江工业大学 | A kind of vehicle queue length detection method based on vehicle tracking |
CN108297794B (en) * | 2018-01-11 | 2022-11-18 | 阿尔派株式会社 | Parking assistance device and travel prediction line display method |
CN109145805B (en) * | 2018-08-15 | 2021-05-14 | 深圳市豪恩汽车电子装备股份有限公司 | Moving target detection method and system under vehicle-mounted environment |
CN109243150A (en) * | 2018-09-30 | 2019-01-18 | 深圳市金豪泰科技有限公司 | A kind of vehicle early warning method and terminal |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101256073A (en) * | 2007-02-28 | 2008-09-03 | 三洋电机株式会社 | Apparatus and method for monitoring a vehicle's surroundings |
CN102774325A (en) * | 2012-07-31 | 2012-11-14 | 西安交通大学 | Rearview reversing auxiliary system and method for forming rearview obstacle images |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2004114879A (en) * | 2002-09-27 | 2004-04-15 | Clarion Co Ltd | Parking assisting device, and image display device |
DE102005023177A1 (en) * | 2005-05-19 | 2006-11-30 | Robert Bosch Gmbh | Method for driver assistance |
EP3594853A3 (en) * | 2007-05-03 | 2020-04-08 | Sony Deutschland GmbH | Method for detecting moving objects in a blind spot region of a vehicle and blind spot detection device |
DE102010042063B4 (en) * | 2010-10-06 | 2021-10-28 | Robert Bosch Gmbh | Method and device for determining processed image data about the surroundings of a vehicle |
CN103150748B (en) * | 2013-03-18 | 2017-03-22 | 大连慈航电子有限公司 | Reversing image three-dimensional (3D) scene reconstruction method and system |
-
2013
- 2013-09-27 CN CN201310451913.0A patent/CN104512329B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101256073A (en) * | 2007-02-28 | 2008-09-03 | 三洋电机株式会社 | Apparatus and method for monitoring a vehicle's surroundings |
CN102774325A (en) * | 2012-07-31 | 2012-11-14 | 西安交通大学 | Rearview reversing auxiliary system and method for forming rearview obstacle images |
Also Published As
Publication number | Publication date |
---|---|
CN104512329A (en) | 2015-04-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104512329B (en) | Move backward safe householder methods, device and automobile | |
CN108725440B (en) | Forward collision control method and apparatus, electronic device, program, and medium | |
CN104517111B (en) | Method for detecting lane lines, system, lane departure warning method and system | |
Chan et al. | Vehicle detection under various lighting conditions by incorporating particle filter | |
CN110738121A (en) | front vehicle detection method and detection system | |
Yousaf et al. | Comparative analysis of automatic vehicle classification techniques: a survey | |
WO2019026785A1 (en) | Attached object detection device, and vehicle system provided with same | |
CN111198371A (en) | Forward-looking obstacle detection system | |
CN107274678B (en) | A kind of night vehicle flowrate and model recognizing method based on Kinect | |
AU2013316125B2 (en) | Backward movement indicator apparatus for a vehicle | |
CN105930787A (en) | Vehicle door opening early-warning method | |
CN109703460A (en) | The complex scene adaptive vehicle collision warning device and method for early warning of multi-cam | |
CN104881645A (en) | Vehicle front target detection method based on characteristic-point mutual information content and optical flow method | |
CN102496281A (en) | Vehicle red-light violation detection method based on combination of tracking and virtual loop | |
CN110298300A (en) | A method of detection vehicle violation crimping | |
CN109409353A (en) | Vehicle whistle monitoring method and system based on DCNN target identification | |
CN1606033A (en) | Weak target detecting and tracking method in infrared image sequence | |
CN106570487A (en) | Method and device for predicting collision between objects | |
Kraus et al. | The radar ghost dataset–an evaluation of ghost objects in automotive radar data | |
CN115100618B (en) | Multi-source heterogeneous perception information multi-level fusion characterization and target identification method | |
Sonoda et al. | Moving objects detection at an intersection by sequential background extraction | |
CN202885814U (en) | Monocular-vision-based real-time barrier detecting device in complex road environment | |
CN107256382A (en) | Virtual bumper control method and system based on image recognition | |
Frías-Velázquez et al. | A mathematical morphology-based approach for vehicle detection in road tunnels | |
Purnama et al. | Real time vehicle counter system for intelligent transportation system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |