CN107784042A - A kind of object matching method and device - Google Patents
A kind of object matching method and device Download PDFInfo
- Publication number
- CN107784042A CN107784042A CN201610799308.6A CN201610799308A CN107784042A CN 107784042 A CN107784042 A CN 107784042A CN 201610799308 A CN201610799308 A CN 201610799308A CN 107784042 A CN107784042 A CN 107784042A
- Authority
- CN
- China
- Prior art keywords
- traffic sign
- road
- matched
- road signs
- candidate roads
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/53—Querying
- G06F16/532—Query formulation, e.g. graphical querying
Abstract
This application discloses a kind of object matching method and device, the application according to road signs to be matched positional information and place road, candidate roads traffic sign is chosen in database, according to the type of the Traffic Sign Images of road signs to be matched, same type candidate roads traffic sign is determined in candidate roads traffic sign, finally according to the similarity of Traffic Sign Images, the object matched with road signs to be matched is determined in same type candidate roads traffic sign.If matching object is not present in road signs to be matched in database, the road signs to be matched are then not present in representation database, and then database can be updated using the road signs to be matched, or by manually to being audited in the absence of the road signs to be matched of matching object, compared to prior art, reduce manual resource and save the time.
Description
Technical field
The application is related to technical field of map data processing, more specifically to a kind of object matching method and device.
Background technology
Road signs are warning, forbid, limit, indicate road user.Road signs are typically with mark
Form be arranged on the sign board of road surface both sides, or be rendered directly on road surface.
The information of each road signs is stored with known road signs database, information includes traffic sign
Image, positional information and place road.With the renewal of road network facility, the traffic sign on road also updates therewith.In order to
The information for ensureing to store in database is newest, it is necessary to which regular be updated to database.Existing common update scheme master
If Usage data collection car carries out picture or video capture, and picture or video by operating personnel to shooting on road surface
It is identified, each Traffic Sign Images being identified out, and determines the positional information and place road of each traffic sign.Enter one
Step, manually audits to aforementioned information, judges to whether there is identical information in database, when it is determined that being not present, before utilization
Information is stated to be updated database.
Obviously, the mode of existing manual work needs to expend a large amount of human resources, and loses time.
The content of the invention
In view of this, this application provides a kind of object matching method and device, for solving existing artificial cross-check information
Mode present in labor intensive resource and the problem of lose time.
To achieve these goals, it is proposed that scheme it is as follows：
A kind of object matching method, including：
According to the positional information of road signs to be matched and place road, in known road signs database
Search and distance identical with the place road of the road signs to be matched and meet that the candidate roads of setpoint distance condition are handed over
Logical mark；
In the candidate roads traffic sign, it is determined that with the Traffic Sign Images of the road signs to be matched
Type identical same type candidate roads traffic sign；
The Traffic Sign Images of each same type candidate roads traffic sign are calculated, with the road traffic mark to be matched
The similarity of the Traffic Sign Images of will；
According to the Similarity Measure result of each same type candidate roads traffic sign, it is determined that with the road to be matched
The same type candidate roads traffic sign of traffic sign matching.
A kind of object matching device, including：
Candidate roads traffic sign searching unit, for the positional information according to road signs to be matched and place road
Road, and distance identical with the place road of the road signs to be matched is searched in known road signs database
Meet the candidate roads traffic sign of setpoint distance condition；
Same type candidate roads traffic sign determining unit, in the candidate roads traffic sign, it is determined that and institute
State the type identical same type candidate roads traffic sign of the Traffic Sign Images of road signs to be matched；
Similarity calculated, for calculating the Traffic Sign Images of each same type candidate roads traffic sign, with
The similarity of the Traffic Sign Images of the road signs to be matched；
Matching result determining unit, for the Similarity Measure knot according to each same type candidate roads traffic sign
Fruit, it is determined that the same type candidate roads traffic sign matched with the road signs to be matched.
It can be seen from the above technical scheme that the object matching method that the embodiment of the present application provides, according to road to be matched
The positional information and place road of road traffic sign, searched and the road to be matched in known road signs database
The place road of traffic sign is identical and distance meets the candidate roads traffic sign of setpoint distance condition, in the candidate roads
In traffic sign, it is determined that the type identical same type candidate road with the Traffic Sign Images of the road signs to be matched
Road traffic sign, the Traffic Sign Images of each same type candidate roads Traffic Sign Images are calculated, with the road to be matched
The similarity of the Traffic Sign Images of road traffic sign, according to the Similarity Measure of each same type candidate roads traffic sign
As a result, it is determined that the same type candidate roads traffic sign matched with the road signs to be matched.As can be seen here, the application
According to the positional information and place road of road signs to be matched, candidate roads traffic sign is chosen in database, is entered
One step determines same type according to the type of the Traffic Sign Images of road signs to be matched in candidate roads traffic sign
Candidate roads traffic sign, finally according to the similarity of Traffic Sign Images, determined in same type candidate roads traffic sign
The object matched with road signs to be matched.If matching object in database be present in road signs to be matched,
The road signs to be matched in representation database then be present, if road signs to be matched in database not
In the presence of matching object, then the road signs to be matched are not present in representation database, and then the road to be matched can be utilized
Road traffic sign is updated to database, or by manually to being carried out in the absence of the road signs to be matched of matching object
Examination ＆ verification, compared to prior art, greatly reduces manual resource and saves the time.
Brief description of the drawings
, below will be to embodiment or existing in order to illustrate more clearly of the embodiment of the present application or technical scheme of the prior art
There is the required accompanying drawing used in technology description to be briefly described, it should be apparent that, drawings in the following description are only this
The embodiment of application, for those of ordinary skill in the art, on the premise of not paying creative work, can also basis
The accompanying drawing of offer obtains other accompanying drawings.
Fig. 1 is a kind of object matching method flow diagram disclosed in the embodiment of the present application；
Fig. 2 is a kind of method flow diagram for searching candidate roads traffic sign disclosed in the embodiment of the present application；
Fig. 3 is a kind of target road traffic sign positional information calibration method flow chart disclosed in the embodiment of the present application；
Fig. 4 is another object matching method flow diagram disclosed in the embodiment of the present application；
Fig. 5 is a kind of object matching apparatus structure schematic diagram disclosed in the embodiment of the present application.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present application, the technical scheme in the embodiment of the present application is carried out clear, complete
Site preparation describes, it is clear that described embodiment is only some embodiments of the present application, rather than whole embodiments.It is based on
Embodiment in the application, those of ordinary skill in the art are obtained every other under the premise of creative work is not made
Embodiment, belong to the scope of the application protection.
The application is deposited to solve the existing processing mode manually checked each road signs of identification
, labor intensive resource, the problem of losing time, it is proposed that a kind of object matching method, searched with realizing in database
The road signs matched with the road signs identified, if matching road traffic mark in database be present
Will, then the road signs identified can be abandoned, if matching road signs are not present in database,
Database can be then updated using the road signs identified, or the road signs that this is identified
Staff is showed, it is audited by staff.Compared to prior art manually one by one to the road traffic of identification
The mode that mark is audited, the application greatly reduce manual resource and save the time.
Next the object matching method of the application is introduced, referring to Fig. 1, Fig. 1 is disclosed in the embodiment of the present application
A kind of object matching method flow diagram.
As shown in figure 1, this method includes：
Step S100, according to the positional information of road signs to be matched and place road, in known road traffic mark
And distance identical with the place road of the road signs to be matched is searched in will database and meets setpoint distance condition
Candidate information point；
Specifically, road signs to be matched include Traffic Sign Images, positional information and place road.Wherein, position
Confidence breath is the geographical position of traffic sign, can be represented by GPS information, and place road is the road letter where traffic sign
Breath, such as road name, numbering.
In this step, in each road signs stored in known road signs database, search and treat
Road where matching road signs is identical, and distance meets the candidate roads traffic sign of setpoint distance condition.
Wherein, setpoint distance condition can include, and the positional distance of two road signs is no more than setpoint distance threshold
Value.
Step S110, in the candidate roads traffic sign, it is determined that the traffic with the road signs to be matched
The type identical same type candidate roads traffic sign of sign image；
Specifically, in the candidate roads traffic sign found in previous step, it is possible to can exist and road to be matched
The different types of Traffic Sign Images of the Traffic Sign Images of road traffic sign.For such Traffic Sign Images, it can not
Can match with road signs to be matched, therefore, in this step from candidate roads traffic sign, determine with it is to be matched
The type identical same type candidate roads traffic sign of the Traffic Sign Images of road signs.
Step S120, the Traffic Sign Images of each same type candidate roads traffic sign are calculated, it is and described to be matched
The similarity of the Traffic Sign Images of road signs；
Specifically, for the same type candidate roads traffic sign obtained in previous step, its Traffic Sign Images is calculated
With the similarity of the Traffic Sign Images of road signs to be matched.
Similarity calculating method may be referred to existing image similarity calculation method, not be described in detail herein.
Step S130, according to the Similarity Measure result of each same type candidate roads traffic sign, it is determined that with it is described
The same type candidate roads traffic sign of road signs matching to be matched.
Specifically, by setting matching condition, can be determined and road traffic mark to be matched according to Similarity Measure result
The same type candidate roads traffic sign of will matching.Certainly, if the Similarity Measure of each same type candidate roads traffic sign
As a result setting matching condition is not met, then can determine that the object matched with road signs to be matched is not present.
As can be seen here, the application according to road signs to be matched positional information and place road, in database
Candidate roads traffic sign is chosen, further according to the type of the Traffic Sign Images of road signs to be matched, in candidate
Same type candidate roads traffic sign is determined in road signs, finally according to the similarity of Traffic Sign Images, similar
The object matched with road signs to be matched is determined in type candidate roads traffic sign.If road signs to be matched
Matching object in database be present, then the road signs to be matched in representation database be present, if to be matched
Matching object is not present in road signs in database, then the road traffic mark to be matched is not present in representation database
Will, and then database can be updated using the road signs to be matched, or by manually in the absence of matching pair
The road signs to be matched of elephant are audited, and compared to prior art, are greatly reduced manual resource and are saved the time.
Optionally, before above-described embodiment step S100, the present processes, which can also increase, obtains road to be matched
The process of traffic sign.The process can include：
1st, the road signs of collection are grouped according to place road, by the road signs on same road
It is divided into same group；
Specifically, collecting vehicle collects multiple road signs in the application, and each road signs include institute
In road markings.Based on this, multiple road signs that the application can be according to road where road signs to collection
It is grouped, the road signs on same road is divided into same group.
2nd, a road signs are arbitrarily read from packet as the road signs to be matched.
Specifically, a road signs can be arbitrarily read from packet for each packet, the application, as treating
With road signs.Certainly, in addition the application can also using one packet in each road signs simultaneously as
Road signs to be matched, while multiple road signs to be matched are matched.
In the application another embodiment, to above-mentioned steps S100, believed according to the position of road signs to be matched
Breath and place road, the place road with the road signs to be matched is searched in known road signs database
Identical and distance meets that the process of the candidate roads traffic sign of setpoint distance condition is introduced.
Referring to Fig. 2, Fig. 2 is a kind of method flow diagram for searching candidate roads traffic sign disclosed in the embodiment of the present application.
As shown in Fig. 2 this method includes：
Step S200, searched and the road signs institute to be matched in the known road signs database
In road identical target road traffic sign；
Specifically, knowable to the introduction based on above-mentioned acquisition road signs process to be matched, road traffic mark to be matched
Will packet in read, and in same packet each road signs place road it is identical.That is, road traffic to be matched
The place road of mark is known, and the place road of each road signs is also in known road signs database
It is known.Therefore, the application searched in known road signs database with the road signs to be matched where
Road identical target road traffic sign.
Step S210, the position letter of the road signs in being grouped according to belonging to the road signs to be matched
Breath, is calibrated to the positional information of the target road traffic sign；
Specifically, due to history collection road signs positional information it is possible that deviation, profit in this step
The positional information of road signs in the packet belonging to road signs to be matched, to the target road traffic sign
Positional information calibrated.By calibrating the position deviation of target road traffic sign in unified adjustment database, with after an action of the bowels
It is more accurate during continuous selection candidate roads traffic sign.
Specific calibration process is referred to hereafter related introduction.
Step S220, according to the positional information of the road signs to be matched, the target after positional information calibration
In road signs, search and meet that the candidate roads of setpoint distance condition are handed over the distance of the road signs to be matched
Logical mark.
Specifically, by being handed in database with road identical target road where the road signs to be matched
The positional information of logical mark is calibrated, and in target road traffic sign after calibration, is searched and road traffic mark to be matched
The distance of will meets the candidate roads traffic sign of setpoint distance condition.
The mode for the lookup candidate roads traffic sign introduced in the present embodiment, searched and road to be matched in database
Road identical target road traffic sign where traffic sign, and then in the packet according to belonging to road signs to be matched
The positional information of road signs, the positional information of target road traffic sign is calibrated, and then according to road to be matched
The positional information of road traffic sign, in the target road traffic sign after positional information is calibrated, search and handed over road to be matched
The distance of logical mark meets the candidate roads traffic sign of setpoint distance condition.By the calibration of positional information, lifting is subsequently looked into
Look for the degree of accuracy of candidate roads traffic sign.
Next, to above-mentioned steps S210, the road traffic in the packet according to belonging to the road signs to be matched
The positional information of mark, the process calibrated to the positional information of the target road traffic sign are introduced.
Referring to Fig. 3, Fig. 3 is a kind of target road traffic sign positional information calibration method stream disclosed in the embodiment of the present application
Cheng Tu.
As shown in figure 3, this method includes：
Step S300, from the target road traffic sign, the target road traffic sign positioned at crossing is obtained；
Specifically, the crossing of the road information point crucial as one, there is mark effect.The application hands in target road
The target road traffic sign positioned at crossing is obtained in logical mark, that is, in target road traffic sign, obtains geographical position
For the target road traffic sign of road cross.
Step S310, in the road signs in packet belonging to the road signs to be matched, traffic is searched
The road signs that sign image matches with the Traffic Sign Images of the target road traffic sign positioned at crossing；
After the target road traffic sign positioned at crossing during database is determined, in road signs to be matched
In road signs in affiliated packet, Traffic Sign Images and the target road traffic sign positioned at crossing are searched
The road signs of Traffic Sign Images matching.When it is implemented, images match condition can be set, and then search and meet figure
As the road signs of matching condition.
Explanation is needed further exist for, the Traffic Sign Images of road signs can include original shooting image, the original
Beginning shooting image includes traffic sign and other environmental sceneries., can in this step when searching the Traffic Sign Images of matching
It is more sharp by the traffic sign and environmental scenery in original shooting image being matched according to original shooting image
In the image for searching matching.
Step S320, using the positional information of the road signs found, to the target road positioned at crossing
The positional information of traffic sign is calibrated；
After finding road signs in previous step, believed using the position of the road signs found
Breath, the positional information for being pointed to the target road traffic sign at crossing are calibrated.
Step S330, it is right according to the skew of the target road traffic sign calibration front and back position information positioned at crossing
Other target road traffic in the target road traffic sign in addition to the target road traffic sign positioned at crossing
The positional information of mark is calibrated.
Specifically, the positional information of the target road traffic sign in previous step to being located at crossing in database is carried out
Calibrate, according to the skew that front and back position information should be calibrated positioned at the target road traffic sign at crossing in this step, in same
The positional information of other target road traffic signs on one road is calibrated.
Lift for example, positioned at crossing target road traffic sign positional information calibration before and after offset by 10m southwards, then may be used
It is unified so that this to be located to the positional information of other target road traffic signs on the road of the target road traffic sign at crossing place
Skew 10m to the south.
Pass through other target roads to being located in database on the target road traffic sign and its place road at crossing
The positional information of traffic sign is calibrated, and ensure that the degree of accuracy of the positional information of target road traffic sign, so as to ensure
The candidate target road signs subsequently chosen are more accurate.
In another embodiment of the application, to above-mentioned steps S130, according to each same type candidate roads traffic
The Similarity Measure result of mark, it is determined that the same type candidate roads traffic sign matched with the road signs to be matched
Process be introduced.
The application can preset similarity condition, in the road signs to be matched being calculated and each institute
After the Similarity Measure result for stating same type candidate roads traffic sign, filter out Similarity Measure result and meet that setting is similar
The same type candidate roads traffic sign of degree condition, as the same type candidate road matched with the road signs to be matched
Road traffic sign.
It is understood that setting similarity condition can be setting similarity threshold, what is be calculated is described to be matched
The Similarity Measure result of road signs and each same type candidate roads traffic sign is Similarity value.And then select
Similarity value is taken to exceed the same type candidate roads traffic sign of setting similarity threshold, as road signs to be matched
The object matched somebody with somebody.
It should be noted that the same type candidate road matched with the road signs to be matched determined in this step
The number of road traffic sign may be zero, one or more., can be numerous from database by the application processing mode
Some same type candidate roads traffic signs matched with road signs to be matched are determined in road signs, are worked
Processing only can be identified to several same type candidate roads traffic signs in personnel, also saved significantly compared to prior art
Human resources have been saved, have reduced recognition time.
It is further alternative, it is each in packet belonging to road signs to be matched on the basis of above-described embodiment
Road signs are calculated after the same type candidate roads traffic sign of matching, in order to determine that each road is handed in packet
Lead to the same type candidate roads traffic sign of mark unique match, the application can be determined by bipartite graph, specific embodiment party
Formula is as follows：
1st, the same of matching is calculated in each road signs in being grouped belonging to the road signs to be matched
After type candidate roads traffic sign, matched using each road signs in the packet, and with each road signs
Same type candidate roads traffic sign, build bipartite graph；
Several same type candidate roads traffic signs of matching are calculated in each road signs in packet.No
With road signs matching same type candidate roads traffic sign between may exist it is overlapping, lift for example, road traffic
The same type candidate roads traffic sign for indicating A1 matchings is B1 and B2；The same type candidate roads of road signs A2 matchings
Traffic sign is B2 and B3.As can be seen here, the same type candidate roads traffic sign of each Self Matchings of road signs A1 and A2
In same object be present：B2.
In order to determine the same type candidate roads traffic sign of unique match, the application for each road signs in packet
Bipartite graph can be built, solves the problem using bipartite graph.A vertex set is each road traffic mark in packet in bipartite graph
The set of will composition, another vertex set is same type candidate roads traffic sign set.It is above-mentioned to have calculated road friendship
The Similarity value of logical mark and each same type candidate roads traffic sign, namely in bipartite graph, each road traffic mark out of packet
It is any in the vertex set of will composition to take out a road signs, its with it is any one similar in another summit set in bipartite graph
The Similarity value of type candidate roads traffic sign is all known.Based on this, following step can be performed.
2nd, according to bipartite graph best match algorithm KM algorithms, calculate be grouped in each road signs unique match it is same
Type candidate roads traffic sign.
Specifically, two vertex sets of bipartite graph are respectively defined as X and Y by the application, wherein set X is each in packet
The vertex set of road signs composition, set Y are the vertex set of same type candidate roads traffic sign composition.Set X
In comprising object be Xi (i=1,2,3 ... n)；In set Y comprising object be Yj (j=1,2,3 ... m).Connecting object Xi
Weights with object Yj side are wij, and weight w ij is equal to the object Xi and object Yj of the frontier juncture connection Similarity value.
By KM algorithms, obtain in the case that each object in set X only associates a line, it is all right in set X
As the weights and value on the side of association；Pair in set Y corresponding to the side that object in the selection weights set X maximum with value associates
As the object with the object unique match in set X.
For example thering is object X1 and X2, object X1 to be matched respectively with Y1, Y2 in a packet, Similarity value is respectively 50 Hes
60；X2 is also matched respectively with Y1, Y2, and similarity is respectively 55 and 50；The then principle based on unique association, it is assumed that X1 and Y1
Match somebody with somebody, now X2 and Y2 matchings, obtain weights is 50+50=100 with value；X1 and Y2 matchings are again assumed that, now X2 and Y1 matchings,
Obtain weights is 60+55=115 with value；115 are more than 100, and therefore, X1 and Y2 are matched, and X2 and Y1 matchings are as finally
Matching result.
By the bipartite graph matching method of the present embodiment, can determine and each road signs unique match in packet
Same type candidate roads traffic sign, ensures the degree of accuracy of object matching.
In another embodiment of the application, another object matching method is introduced, referring to Fig. 4.
As shown in figure 4, this method includes：
Step S400, according to the positional information of road signs to be matched and place road, in known road traffic mark
And distance identical with the place road of the road signs to be matched is searched in will database and meets setpoint distance condition
Candidate information point；
Step S410, in the candidate roads traffic sign, it is determined that the traffic with the road signs to be matched
The type identical same type candidate roads traffic sign of sign image；
Step S420, the Traffic Sign Images of each same type candidate roads traffic sign are calculated, it is and described to be matched
The similarity of the Traffic Sign Images of road signs；
Step S430, according to the Similarity Measure result of each same type candidate roads traffic sign, it is determined that with it is described
The same type candidate roads traffic sign of road signs matching to be matched；
Specifically, above-mentioned steps S400-S430 and step S100-S130 in previous embodiment are corresponding respectively, implement in detail
Process is referred to previous embodiment introduction.
Step S440, to the Traffic Sign Images of the road signs to be matched, and its same type candidate of matching
The Traffic Sign Images of road signs carry out color verification；
Specifically, determine each same type candidate roads traffic sign for being matched with road signs to be matched it
Afterwards, the step of further increase carries out color verification to the Traffic Sign Images of road signs.Verified by color, it is determined that
The color of the Traffic Sign Images of same type candidate roads traffic sign, the Traffic Sign Images with road signs to be matched
Color distortion it is whether excessive.
Step S450, color check results are shown that color distortion exceedes the same type candidate road of setpoint color discrepancy threshold
Road traffic sign filters out.
Specifically, if the color of a certain same type candidate roads traffic sign matched with road signs to be matched
Check results show that the color distortion of Traffic Sign Images exceedes setpoint color discrepancy threshold, then it represents that the matching is error hiding,
The same type candidate roads traffic sign is filtered out, namely removes the same type road signs and the road to be matched
Matching relationship between traffic sign.
Compared to above-described embodiment, the process that color verification is carried out to matching result is further increased in the present embodiment,
Verified by color, the excessive matching result of color distortion is filtered out, further increases matching accuracy.
In another embodiment of the application, to above-mentioned steps S120, each same type candidate roads traffic is calculated
The Traffic Sign Images of mark, it is situated between with the process of the similarity of the Traffic Sign Images of the road signs to be matched
Continue.
When calculating similarity, the application is directed to the different type of Traffic Sign Images, using different calculations.
Optionally, the type of Traffic Sign Images includes but is not limited to：Blue party board type, warning and ban type and ground wire
Type.
Wherein, the Traffic Sign Images of blue party board type are mainly, using the rectangle plate of blueness as background, in plate subscript
Knowledge has transport information.
Warn and the Traffic Sign Images of ban type are mainly, the shape such as circle/square/triangle, it is red/white
The colors such as color/green, the sign board of the filling mode such as solid/annulus, such as " no parking " Sign Board.
The Traffic Sign Images of ground wire style are mainly the transport information being plotted on road surface, such as " left turn lane ", " right side
Change trains or buses " etc..
In the present embodiment：
If the 1, the type of the Traffic Sign Images of the road signs to be matched is blue party board type, can use
Dense matching algorithm calculates the Traffic Sign Images of each same type candidate roads traffic sign, is handed over the road to be matched
The similarity of the Traffic Sign Images of logical mark.
Specifically, for blue party board type, because image is bigger, the similitude on Qie Yitiao roads is larger, so, use
Dense matching algorithm calculates similarity.
On this basis, geometric correction means can also be coordinated, geometric correction is carried out to Traffic Sign Images.
If the 2, the type of the Traffic Sign Images of the road signs to be matched is warning and ban type, use
Convolutional neural networks algorithm calculates the Traffic Sign Images of each same type candidate roads traffic sign, with the road to be matched
The similarity of the Traffic Sign Images of road traffic sign.
If the 3rd, the type of the Traffic Sign Images of the road signs to be matched is ground wire style, using color phase
The Traffic Sign Images of each same type candidate roads traffic sign are calculated like degree algorithm, with the road traffic mark to be matched
The similarity of the Traffic Sign Images of will.
Specifically, for ground wire style Traffic Sign Images, because road surface is typically different from the color of traffic sign, therefore
It can select to calculate image similarity using color similarity algorithm.
Certainly, above-described embodiment merely illustrates the type of several Traffic Sign Images, and corresponding Similarity Measure
Mode, in addition those skilled in the art be also an option that other suitable algorithms carry out the calculating of similarities.
Below to the embodiment of the present application provide object matching device be described, object matching device described below with
Above-described object matching method can be mutually to should refer to.
Referring to Fig. 5, Fig. 5 is a kind of object matching apparatus structure schematic diagram disclosed in the embodiment of the present application.
As shown in figure 5, the device includes：
Candidate roads traffic sign searching unit 51, for the positional information according to road signs to be matched and place
Road, in known road signs database search it is identical with the place road of the road signs to be matched and away from
From the candidate roads traffic sign for meeting setpoint distance condition；
Same type candidate roads traffic sign determining unit 52, in the candidate roads traffic sign, it is determined that with
The type identical same type candidate roads traffic sign of the Traffic Sign Images of the road signs to be matched；
Similarity calculated 53, for calculating the Traffic Sign Images of each same type candidate roads traffic sign,
With the similarity of the Traffic Sign Images of the road signs to be matched；
Matching result determining unit 54, for the Similarity Measure knot according to each same type candidate roads traffic sign
Fruit, it is determined that the same type candidate roads traffic sign matched with the road signs to be matched.
The object matching device that the embodiment of the present application provides, positional information and place according to road signs to be matched
Road, candidate roads traffic sign is chosen in database, further according to the traffic indication map of road signs to be matched
The type of picture, same type candidate roads traffic sign is determined in candidate roads traffic sign, finally according to Traffic Sign Images
Similarity, the object that is matched with road signs to be matched is determined in same type candidate roads traffic sign.If treat
Match road signs and matching object in database be present, then the road traffic to be matched in representation database be present
Mark, if matching object is not present in road signs to be matched in database, treated in representation database in the absence of this
Road signs are matched, and then database can be updated using the road signs to be matched, or by artificial
To being audited in the absence of the road signs to be matched of matching object, compared to prior art, people's wage is greatly reduced
Source and save the time.
Optionally, the device of the application can also include：
Grouped element, for the road signs of collection to be grouped according to place road, by same road
Road signs are divided into same group；
Road signs determining unit to be matched, for arbitrarily reading a road signs as institute from packet
State road signs to be matched.
Further alternative, the candidate roads traffic sign searching unit can include：
Target road traffic sign searching unit, in the known road signs database search with it is described
Road identical target road traffic sign where road signs to be matched；
Location calibration unit, for the road signs in the packet according to belonging to the road signs to be matched
Positional information, the positional information of the target road traffic sign is calibrated；
Distance condition searching unit, for the positional information according to the road signs to be matched, in positional information
In target road traffic sign after calibration, search and meet setpoint distance condition with the distance of the road signs to be matched
Candidate roads traffic sign.
Optionally, above-mentioned location calibration unit can include：
Crossing road signs searching unit, for from the target road traffic sign, obtaining positioned at crossing
Target road traffic sign；
Crossing road signs matching unit, for the road in packet belonging to the road signs to be matched
In traffic sign, search Traffic Sign Images and matched with the Traffic Sign Images of the target road traffic sign positioned at crossing
Road signs；
Crossing road signs alignment unit, for the positional information using the road signs found, to institute
Rheme is calibrated in the positional information of the target road traffic sign at crossing；
Other road signs alignment units, before according to the target road traffic sign calibration positioned at crossing
The skew of positional information afterwards, in the target road traffic sign in addition to the target road traffic sign positioned at crossing
The positional informations of other target road traffic signs calibrated.
Optionally, the matching result determining unit can include：
First matching result determination subelement, for according to the road signs to be matched being calculated and each institute
The Similarity Measure result of same type candidate roads traffic sign is stated, Similarity Measure result is filtered out and meets setting similarity bar
The same type candidate roads traffic sign of part, handed over as the same type candidate roads matched with the road signs to be matched
Logical mark.
Further optional, the device of the application can also include：
Bipartite graph establishes unit, for belonging to the road signs to be matched packet in each road signs it is equal
Be calculated after the same type candidate roads traffic sign of matching, using each road signs in the packet, and with it is each
The same type candidate roads traffic sign of road signs matching, builds bipartite graph；
KM computing units, for according to bipartite graph best match algorithm KM algorithms, calculating and each road traffic mark in packet
The same type candidate roads traffic sign of will unique match.
Optionally, the device of the application can also include：
Color verification unit, for the Traffic Sign Images to the road signs to be matched, and its match same
The Traffic Sign Images of type candidate roads traffic sign carry out color verification；
Error hiding filters out unit, for color check results to be shown into color distortion exceedes the same of setpoint color discrepancy threshold
Type candidate roads traffic sign filters out.
Optionally, different according to the type of the Traffic Sign Images of road signs to be matched, the application describes
The alternative construction of similarity calculated is stated, the similarity calculated can include：
First Similarity Measure subelement, if the type of the Traffic Sign Images for the road signs to be matched
For blue party board type, then the traffic indication map of each same type candidate roads traffic sign is calculated using dense matching algorithm
Picture, the similarity with the Traffic Sign Images of the road signs to be matched；
Second Similarity Measure subelement, if the type of the Traffic Sign Images for the road signs to be matched
For warning and ban type, then the traffic of each same type candidate roads traffic sign is calculated using convolutional neural networks algorithm
Sign image, the similarity with the Traffic Sign Images of the road signs to be matched；
Third phase seemingly spends computation subunit, if the type of the Traffic Sign Images for the road signs to be matched
For ground wire style, then the traffic indication map of each same type candidate roads traffic sign is calculated using color similarity algorithm
Picture, the similarity with the Traffic Sign Images of the road signs to be matched.
Finally, it is to be noted that, herein, such as first and second or the like relational terms be used merely to by
One entity or operation make a distinction with another entity or operation, and not necessarily require or imply these entities or operation
Between any this actual relation or order be present.Moreover, term " comprising ", "comprising" or its any other variant meaning
Covering including for nonexcludability, so that process, method, article or equipment including a series of elements not only include that
A little key elements, but also the other element including being not expressly set out, or also include for this process, method, article or
The intrinsic key element of equipment.In the absence of more restrictions, the key element limited by sentence "including a ...", is not arranged
Except other identical element in the process including the key element, method, article or equipment being also present.
Each embodiment is described by the way of progressive in this specification, what each embodiment stressed be and other
The difference of embodiment, between each embodiment identical similar portion mutually referring to.
The foregoing description of the disclosed embodiments, professional and technical personnel in the field are enable to realize or using the application.
A variety of modifications to these embodiments will be apparent for those skilled in the art, as defined herein
General Principle can be realized in other embodiments in the case where not departing from spirit herein or scope.Therefore, the application
The embodiments shown herein is not intended to be limited to, and is to fit to and principles disclosed herein and features of novelty phase one
The most wide scope caused.
Claims (16)
- A kind of 1. object matching method, it is characterised in that including：According to the positional information of road signs to be matched and place road, searched in known road signs database Identical and distance meets the candidate roads traffic mark of setpoint distance condition with the place road of the road signs to be matched Will；In the candidate roads traffic sign, it is determined that the type with the Traffic Sign Images of the road signs to be matched Identical same type candidate roads traffic sign；The Traffic Sign Images of each same type candidate roads traffic sign are calculated, with the road signs to be matched The similarity of Traffic Sign Images；According to the Similarity Measure result of each same type candidate roads traffic sign, it is determined that with the road traffic to be matched The same type candidate roads traffic sign of tag match.
- 2. according to the method for claim 1, it is characterised in that methods described further comprises：The road signs of collection are grouped according to place road, the road signs on same road are divided into together One group；A road signs are arbitrarily read from packet as the road signs to be matched.
- 3. according to the method for claim 2, it is characterised in that the positional information according to road signs to be matched And place road, the place road phase with the road signs to be matched is searched in known road signs database Same and distance meets the candidate roads traffic sign of setpoint distance condition, including：Searched in the known road signs database and road identical where the road signs to be matched Target road traffic sign；The positional information of road signs in the packet according to belonging to the road signs to be matched, to the target track The positional information of road traffic sign is calibrated；According to the positional information of the road signs to be matched, the target road traffic sign after positional information calibration In, search the candidate roads traffic sign for meeting setpoint distance condition with the distance of the road signs to be matched.
- 4. according to the method for claim 3, it is characterised in that described to be divided according to belonging to the road signs to be matched The positional information of road signs in group, the positional information of the target road traffic sign is calibrated, including：From the target road traffic sign, the target road traffic sign positioned at crossing is obtained；Belonging to the road signs to be matched packet in road signs in, search Traffic Sign Images with it is described Positioned at the road signs of the Traffic Sign Images matching of the target road traffic sign at crossing；Using the positional information of the road signs found, to the position of the target road traffic sign positioned at crossing Information is calibrated；The skew of front and back position information is calibrated according to the target road traffic sign positioned at crossing, the target road is handed over The positional information of other target road traffic signs in logical mark in addition to the target road traffic sign positioned at crossing Calibrated.
- 5. according to the method for claim 2, it is characterised in that described according to each same type candidate roads traffic sign Similarity Measure result, it is determined that the same type candidate roads traffic sign matched with the road signs to be matched, bag Include：It is similar to each same type candidate roads traffic sign according to the road signs to be matched being calculated Result of calculation is spent, filters out the same type candidate roads traffic sign that Similarity Measure result meets setting similarity condition, is made For the same type candidate roads traffic sign matched with the road signs to be matched.
- 6. according to the method for claim 5, it is characterised in that this method also includes：The same type time of matching is calculated in each road signs in being grouped belonging to the road signs to be matched After selecting road signs, using each road signs in the packet, and match with each road signs similar Type candidate roads traffic sign, build bipartite graph；According to bipartite graph best match algorithm KM algorithms, calculate and waited with the same type of each road signs unique match in packet Select road signs.
- 7. according to the method described in claim any one of 1-6, it is characterised in that also include：To the Traffic Sign Images of the road signs to be matched, and its same type candidate roads traffic sign matched Traffic Sign Images carry out color verification；Color check results are shown that color distortion exceedes the same type candidate roads traffic sign filter of setpoint color discrepancy threshold Remove.
- 8. according to the method described in claim any one of 1-6, it is characterised in that described to calculate each same type candidate roads The Traffic Sign Images of traffic sign, the similarity with the Traffic Sign Images of the road signs to be matched, including：If the type of the Traffic Sign Images of the road signs to be matched is blue party board type, calculated using dense matching Method calculates the Traffic Sign Images of each same type candidate roads traffic sign, the friendship with the road signs to be matched The similarity of logical sign image；If the type of the Traffic Sign Images of the road signs to be matched is warning and ban type, using convolution god The Traffic Sign Images of each same type candidate roads traffic sign are calculated through network algorithm, with the road traffic to be matched The similarity of the Traffic Sign Images of mark；If the type of the Traffic Sign Images of the road signs to be matched is ground wire style, calculated using color similarity Method calculates the Traffic Sign Images of each same type candidate roads traffic sign, the friendship with the road signs to be matched The similarity of logical sign image.
- A kind of 9. object matching device, it is characterised in that including：Candidate roads traffic sign searching unit, for the positional information according to road signs to be matched and place road, Lookup is identical with the place road of the road signs to be matched and apart from full in known road signs database The candidate roads traffic sign of sufficient setpoint distance condition；Same type candidate roads traffic sign determining unit, in the candidate roads traffic sign, it is determined that being treated with described Match the type identical same type candidate roads traffic sign of the Traffic Sign Images of road signs；Similarity calculated, it is and described for calculating the Traffic Sign Images of each same type candidate roads traffic sign The similarity of the Traffic Sign Images of road signs to be matched；Matching result determining unit, for the Similarity Measure result according to each same type candidate roads traffic sign, really The fixed same type candidate roads traffic sign matched with the road signs to be matched.
- 10. device according to claim 9, it is characterised in that also include：Grouped element, for the road signs of collection to be grouped according to place road, by the road on same road Traffic sign is divided into same group；Road signs determining unit to be matched, for being treated described in one road signs conduct of any reading from packet Match road signs.
- 11. device according to claim 10, it is characterised in that the candidate roads traffic sign searching unit includes：Target road traffic sign searching unit, treated with described for being searched in the known road signs database With road identical target road traffic sign where road signs；Location calibration unit, the position for the road signs in the packet according to belonging to the road signs to be matched Information, the positional information of the target road traffic sign is calibrated；Distance condition searching unit, for the positional information according to the road signs to be matched, calibrated in positional information In target road traffic sign afterwards, the time for meeting setpoint distance condition with the distance of the road signs to be matched is searched Select road signs.
- 12. device according to claim 11, it is characterised in that the location calibration unit includes：Crossing road signs searching unit, for from the target road traffic sign, obtaining the target positioned at crossing Road signs；Crossing road signs matching unit, for the road traffic in packet belonging to the road signs to be matched In mark, the road that Traffic Sign Images match with the Traffic Sign Images of the target road traffic sign positioned at crossing is searched Road traffic sign；Crossing road signs alignment unit, for the positional information using the road signs found, to institute's rheme Calibrated in the positional information of the target road traffic sign at crossing；Other road signs alignment units, for calibrating anteroposterior position according to the target road traffic sign positioned at crossing The skew of confidence breath, to its in the target road traffic sign in addition to the target road traffic sign positioned at crossing The positional information of its target road traffic sign is calibrated.
- 13. device according to claim 10, it is characterised in that the matching result determining unit includes：First matching result determination subelement, for according to the road signs to be matched that are calculated with it is each described same The Similarity Measure result of type candidate roads traffic sign, filter out Similarity Measure result and meet setting similarity condition Same type candidate roads traffic sign, as the same type candidate roads traffic mark matched with the road signs to be matched Will.
- 14. device according to claim 13, it is characterised in that also include：Bipartite graph establishes unit, for belonging to the road signs to be matched packet in each road signs calculate After the same type candidate roads traffic sign matched, using each road signs in the packet, and with each road The same type candidate roads traffic sign of traffic sign matching, builds bipartite graph；KM computing units, for according to bipartite graph best match algorithm KM algorithms, calculating with each road signs in packet only The same type candidate roads traffic sign of one matching.
- 15. according to the device described in claim any one of 9-14, it is characterised in that also include：Color verification unit, for the Traffic Sign Images to the road signs to be matched, and its same type of matching The Traffic Sign Images of candidate roads traffic sign carry out color verification；Error hiding filters out unit, for color check results to be shown into color distortion exceedes the same type of setpoint color discrepancy threshold Candidate roads traffic sign filters out.
- 16. according to the device described in claim any one of 9-14, it is characterised in that the similarity calculated includes：First Similarity Measure subelement, if the type for the Traffic Sign Images of the road signs to be matched is indigo plant Square board type, then the Traffic Sign Images of each same type candidate roads traffic sign are calculated using dense matching algorithm, with The similarity of the Traffic Sign Images of the road signs to be matched；Second Similarity Measure subelement, if the type for the Traffic Sign Images of the road signs to be matched is police Show with ban type, then the traffic sign of each same type candidate roads traffic sign is calculated using convolutional neural networks algorithm Image, the similarity with the Traffic Sign Images of the road signs to be matched；Third phase seemingly spends computation subunit, if the type for the Traffic Sign Images of the road signs to be matched is ground Line type, then the Traffic Sign Images of each same type candidate roads traffic sign are calculated using color similarity algorithm, with The similarity of the Traffic Sign Images of the road signs to be matched.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610799308.6A CN107784042B (en) | 2016-08-31 | 2016-08-31 | Object matching method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610799308.6A CN107784042B (en) | 2016-08-31 | 2016-08-31 | Object matching method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107784042A true CN107784042A (en) | 2018-03-09 |
| CN107784042B CN107784042B (en) | 2021-02-05 |
Family
ID=61451839
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610799308.6A Active CN107784042B (en) | 2016-08-31 | 2016-08-31 | Object matching method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107784042B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108922201A (en) * | 2018-07-13 | 2018-11-30 | 江苏海事职业技术学院 | A kind of traffic lights speech prompting system |
| CN109949692A (en) * | 2019-03-27 | 2019-06-28 | 腾讯大地通途（北京）科技有限公司 | Road network method, apparatus, computer equipment and storage medium |
| CN110348199A (en) * | 2019-07-17 | 2019-10-18 | 上海图趣信息科技有限公司 | Picture labeling system, method, storage medium and picture recognition self learning system based on identifying code |
| CN110751133A (en) * | 2019-12-03 | 2020-02-04 | 清华大学苏州汽车研究院（吴江） | Road traffic sign prompt system |
| CN111951466A (en) * | 2020-08-24 | 2020-11-17 | 宿州赛尔沃德物联网科技有限公司 | Method for realizing safe circulation and automatic vending of medicines |
| CN114199228A (en) * | 2021-12-08 | 2022-03-18 | 北京百度网讯科技有限公司 | Map data generation method and device, electronic equipment and storage medium |
| CN115827925A (en) * | 2023-02-21 | 2023-03-21 | 中国第一汽车股份有限公司 | Target association method and device, electronic equipment and storage medium |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11887379B2 (en) | 2021-08-17 | 2024-01-30 | Robert Bosch Gmbh | Road sign content prediction and search in smart data management for training machine learning model |
| US11978264B2 (en) | 2021-08-17 | 2024-05-07 | Robert Bosch Gmbh | Scalable road sign knowledge graph construction with machine learning and human in the loop |
| DE102022125693A1 (en) | 2022-10-05 | 2024-04-11 | Bayerische Motoren Werke Aktiengesellschaft | Method and device for recognizing traffic signs in the environment of a vehicle taking into account a color distribution and vehicle |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104182760A (en) * | 2014-04-16 | 2014-12-03 | 南京通用电器有限公司 | Traffic information reading method based on coding traffic sign |
| US20160117562A1 (en) * | 2014-10-23 | 2016-04-28 | Hyundai Mobis Co., Ltd. | Traffic sign recognizing apparatus and operating method thereof |
| CN105719499A (en) * | 2016-04-21 | 2016-06-29 | 百度在线网络技术（北京）有限公司 | Traffic sign recognition test method and device |
| CN105788323A (en) * | 2014-12-24 | 2016-07-20 | 华创车电技术中心股份有限公司 | Road speed limit detecting system |
-
2016
- 2016-08-31 CN CN201610799308.6A patent/CN107784042B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104182760A (en) * | 2014-04-16 | 2014-12-03 | 南京通用电器有限公司 | Traffic information reading method based on coding traffic sign |
| US20160117562A1 (en) * | 2014-10-23 | 2016-04-28 | Hyundai Mobis Co., Ltd. | Traffic sign recognizing apparatus and operating method thereof |
| CN105788323A (en) * | 2014-12-24 | 2016-07-20 | 华创车电技术中心股份有限公司 | Road speed limit detecting system |
| CN105719499A (en) * | 2016-04-21 | 2016-06-29 | 百度在线网络技术（北京）有限公司 | Traffic sign recognition test method and device |
Non-Patent Citations (1)
| Title |
|---|
| 彭岳军: "一种基于SURF的交通标志识别算法", 《信息技术》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108922201A (en) * | 2018-07-13 | 2018-11-30 | 江苏海事职业技术学院 | A kind of traffic lights speech prompting system |
| CN109949692A (en) * | 2019-03-27 | 2019-06-28 | 腾讯大地通途（北京）科技有限公司 | Road network method, apparatus, computer equipment and storage medium |
| CN109949692B (en) * | 2019-03-27 | 2021-03-26 | 腾讯大地通途（北京）科技有限公司 | Road network matching method and device, computer equipment and storage medium |
| CN110348199A (en) * | 2019-07-17 | 2019-10-18 | 上海图趣信息科技有限公司 | Picture labeling system, method, storage medium and picture recognition self learning system based on identifying code |
| CN110751133A (en) * | 2019-12-03 | 2020-02-04 | 清华大学苏州汽车研究院（吴江） | Road traffic sign prompt system |
| CN110751133B (en) * | 2019-12-03 | 2023-08-18 | 清华大学苏州汽车研究院（吴江） | Road traffic sign prompting system |
| CN111951466A (en) * | 2020-08-24 | 2020-11-17 | 宿州赛尔沃德物联网科技有限公司 | Method for realizing safe circulation and automatic vending of medicines |
| CN114199228A (en) * | 2021-12-08 | 2022-03-18 | 北京百度网讯科技有限公司 | Map data generation method and device, electronic equipment and storage medium |
| CN115827925A (en) * | 2023-02-21 | 2023-03-21 | 中国第一汽车股份有限公司 | Target association method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107784042B (en) | 2021-02-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107784042A (en) | A kind of object matching method and device | |
| JP5057183B2 (en) | Reference data generation system and position positioning system for landscape matching | |
| CN103473774B (en) | A kind of vehicle positioning method based on pavement image characteristic matching | |
| JP5435306B2 (en) | Image processing system and positioning system | |
| JP5505723B2 (en) | Image processing system and positioning system | |
| CN101275841B (en) | Feature information collecting apparatus and feature information collecting method | |
| JP5062498B2 (en) | Reference data generation system and position positioning system for landscape matching | |
| JP5057184B2 (en) | Image processing system and vehicle control system | |
| CN110148196A (en) | A kind of image processing method, device and relevant device | |
| CN104508719B (en) | Drive assist system and driving assistance method | |
| JP5168601B2 (en) | Own vehicle position recognition system | |
| JP4970195B2 (en) | Person tracking system, person tracking apparatus, and person tracking program | |
| CN107341523A (en) | Express delivery list information identifying method and system based on deep learning | |
| CN106462727A (en) | Systems and methods for lane end recognition | |
| CN108362293A (en) | A kind of track of vehicle matching process based on key point technology | |
| CN108428254A (en) | The construction method and device of three-dimensional map | |
| WO2021056341A1 (en) | Lane line fusion method, lane line fusion apparatus, vehicle, and storage medium | |
| CN109447069A (en) | Collecting vehicle information recognition methods and system towards intelligent terminal | |
| CN106558051A (en) | A kind of improved method for detecting road from single image | |
| JP2011215055A (en) | One's own vehicle position detection system using scenery image recognition | |
| CN110135323A (en) | Image labeling method, device, system and storage medium | |
| CN109785637A (en) | The assay method and device of rule-breaking vehicle | |
| JP2011215052A (en) | Own-vehicle position detection system using scenic image recognition | |
| JP2015194397A (en) | Vehicle location detection device, vehicle location detection method, vehicle location detection computer program and vehicle location detection system | |
| JP4292250B2 (en) | Road environment recognition method and road environment recognition device |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200508 Address after: 310052 room 508, floor 5, building 4, No. 699, Wangshang Road, Changhe street, Binjiang District, Hangzhou City, Zhejiang Province Applicant after: Alibaba (China) Co.,Ltd. Address before: 102200, No. 8, No., Changsheng Road, Changping District science and Technology Park, Beijing, China. 1-5 Applicant before: AUTONAVI SOFTWARE Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |