CN106372255B - A kind of recognition methods of street view image and device - Google Patents
A kind of recognition methods of street view image and device Download PDFInfo
- Publication number
- CN106372255B CN106372255B CN201610865626.8A CN201610865626A CN106372255B CN 106372255 B CN106372255 B CN 106372255B CN 201610865626 A CN201610865626 A CN 201610865626A CN 106372255 B CN106372255 B CN 106372255B
- Authority
- CN
- China
- Prior art keywords
- image
- target
- projected position
- image set
- projected
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- 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
- G06F16/51—Indexing; Data structures therefor; Storage structures
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Remote Sensing (AREA)
- Processing Or Creating Images (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention discloses a kind of recognition methods of street view image and device, wherein this method, comprising: the image for determining target road and matching with target road；According to the position of the collection point of each described image, projected position of the collection point in the target road is determined, and each described image is screened according to projected position.The embodiment of the present invention can identify a large amount of repeated acquisitions with path adaptation street view image, and the street view image to match with road is accurately shown in panoramic table, improves efficiency when street view image and path adaptation.
Description
Technical field
The present embodiments relate to the recognition methods of Cartographic Technique more particularly to a kind of street view image and devices.
Background technique
With the rapid development of economy, the diversification of the vehicles provides quick, efficient trip mode for people, because
And the demand of a large amount of point of interest (Point of Interest, POI) search and traffic path planning is produced, electronic map
Significant progress is obtained, street view service is also come into being.User may browse through the streetscape with path adaptation by electronic map
Image.
Wherein, it needs to carry out Image Acquisition to city by mapping vehicle with the street view image of path adaptation, but is adopting
During collecting image, since mapping vehicle is different, acquisition time is different or some other factors lead to the streetscape figure acquired
There are the images of a large amount of repeated acquisition as in.In the prior art, the image of this repeated acquisition can not be identified, is reduced
Efficiency when street view image and path adaptation.
Summary of the invention
The present invention provides recognition methods and the device of a kind of street view image, to realize identify a large amount of repeated acquisitions and road
Matched street view image accurately shows the street view image to match with road in panoramic table, improves street view image and road
Efficiency when matching.
In a first aspect, the embodiment of the invention provides a kind of recognition methods of street view image, comprising:
The image for determining target road and matching with target road；
According to the position of the collection point of each described image, projected position of the collection point in the target road is determined, and
Each described image is screened according to projected position.
Second aspect, the embodiment of the invention also provides a kind of identification devices of street view image, comprising:
Determining module, the image for determining target road and matching with target road；
Screening module determines collection point in the target road for the position according to the collection point of each described image
Projected position, and each described image is screened according to projected position.
Technical solution provided in an embodiment of the present invention, by the determining Image Acquisition point to match with target road in target
Projected position on road, and screened according to projected position pair and the image that target road matches, it can identify a large amount of
The street view image with path adaptation of repeated acquisition accurately shows the street view image to match with road in panoramic table, mentions
Efficiency when high street view image and path adaptation.
Detailed description of the invention
Fig. 1 is a kind of recognition methods flow chart for street view image that the embodiment of the present invention one provides；；
Fig. 2 a is a kind of recognition methods flow chart of street view image provided by Embodiment 2 of the present invention；
Fig. 2 b is the collection point position view of the image provided by Embodiment 2 of the present invention that matches with target road；
After Fig. 2 c is the image duplicate removal processing provided by Embodiment 2 of the present invention that matches with target road, collection point position shows
It is intended to；
Fig. 3 a is a kind of recognition methods flow chart for street view image that the embodiment of the present invention three provides；
Fig. 3 b is the collection point position view for the image that matches with target road that the embodiment of the present invention three provides；
Fig. 3 c is the matching after image duplicate removal processing with target road of providing of the embodiment of the present invention three, the first collection point
Position view；
Fig. 3 d is the matching after image duplicate removal processing with target road of providing of the embodiment of the present invention three, second of collection point
Position view；
Fig. 4 is the identification device structural block diagram for the street view image that the embodiment of the present invention four provides.
Specific embodiment
The present invention is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched
The specific embodiment stated is used only for explaining the present invention rather than limiting the invention.It also should be noted that in order to just
Only the parts related to the present invention are shown in description, attached drawing rather than entire infrastructure.
Embodiment one
Fig. 1 is a kind of recognition methods flow chart for street view image that the embodiment of the present invention one provides, and the method is by streetscape
The identification device of image executes, and described device is executed by software and/or hardware, described device is generally disposed at server
In.As shown in Figure 1, technical solution provided in this embodiment is specific as follows:
S110: the image for determining target road and matching with target road.
In the present embodiment, the information for being stored with each road in the database, the figure to match with each road information
As, the acquisition time of each image and the position of collection point, other data can also be stored in database.Wherein, road
The information on road includes the identification information of road, road GEO data etc..The identification information of road can be the number information to road
Or title of road etc..
In the present embodiment, the position of collection point can be determined by latitude and longitude coordinates, can also be based on target road
The coordinate system of foundation is determined.Wherein, the coordinate system established based on target road can be with the center line of target road, side
Boundary line or other straight lines parallel with road are X-axis, on same road surface, and the straight line vertical with X-axis is Y-axis with X-axis.
In the present embodiment, when choosing target road in the database, due in the database, the title of road, road
Number and the image to match with road form mapping relations, therefore when target road determines, match with target road
Image also determines.It wherein, can be according to the row of the identification information of each road in database for the method for determination of target road
Sequence is successively determined.Wherein, the identification information of each road information of database is ranked up according to default rule.It is preset
Rule can be position, the weighted value of road etc. where road.The method of determination of target road is also possible to pass through people
The mode that work is controlled as desired by the identification information that target road is manually entered, and then determines target road.Work as mesh
After mark road determines, because of the mapping relations of target road and image, the image to match with target road can also be carried out really
It is fixed.
S120: according to the position of the collection point of each described image, projection position of the collection point in the target road is determined
It sets, and each described image is screened according to projected position.
In the present embodiment, the position when position of the collection point of image refers to acquisition image where Operation Van.Work as operation
When acquiring the image to match with target road in target road, Operation Van will appear left-hand rotation in target road, turns right vehicle
Situation, i.e. Operation Van is not to maintain the same moving direction, i.e. shifting of the Operation Van in target road in target road
Dynamic route is not straight line.Therefore the position where Operation Van is indicated in the form of two-dimensional coordinate.Therefore, image is adopted
The position of collection point is indicated in the form of two-dimensional coordinate.
In the present embodiment, optionally, the position of the collection point according to each described image, determines collection point described
Projected position in target road, and carrying out screening to each described image according to projected position includes: by the collection point of each image
It is projected to target road, and each image is screened according to the distance between projected position.Wherein, adopting each image
Collection point, which carries out projection to target road, can be the collection point of each image to the center line of target road, boundary line or and mesh
The parallel straight line of mark road is projected.
Alternatively, the position of the collection point according to each described image, determines throwing of the collection point in the target road
Shadow position, and carrying out screening to each described image according to projected position includes: the acquisition from each described image according to collecting vehicle
Time is divided at least two image collections, and is determined according to the weighted value of at least two image collection and choose target image
Set；The collection point position of the image in each image collection is projected in the target road, to determine projected position；It will
The projected position of each image in non-object image set, is compared with the image projection positions in target image set one by one；
The image in non-targeted combination is inserted into the target image set according to projected position based on comparative result.Wherein, scheme
The weighted value that image set closes can be determined by the acquisition time of image, the position of the collection point of image or other factors.Pass through
This method can according to need the image retained in target image set and be kept away based on the image version in target image set
The phenomenon that image for having exempted to occur during different images version and path adaptation jumps.
It wherein, can be according to the size of weighted value when including that at least two image subsets are closed in non-object image set
Determine the comparative sequence of the projected position and image projection positions in target image set of image in image subset conjunction.Pass through the party
Method, can identify the image of a large amount of repeated acquisitions, and can according to the needs of users according to the significance level of image version successively
Retain different image versions.
A kind of recognition methods of street view image provided in this embodiment is adopted by determining with the image that target road matches
Projected position of the collection point in target road, and screened according to projected position pair with the image that target road matches, energy
It enough identifies a large amount of repeated acquisitions with path adaptation street view image, the street to match with road is accurately shown in panoramic table
Scape image improves efficiency when street view image and path adaptation.
Embodiment two
Fig. 2 a is a kind of recognition methods flow chart of street view image provided by Embodiment 2 of the present invention, in above-described embodiment one
On the basis of, optionally, according to the position of the collection point of each described image, determine projection of the collection point in the target road
Position, and screening is carried out to each described image according to projected position and includes:
According to the position of the collection point of each described image, projected position of the collection point in the target road is determined；
The image that the distance of projected position is less than setting overlapping threshold value is subjected to duplicate removal processing.
Duplicate removal processing, Neng Goushi are carried out by the image that the distance of projected position is less than setting overlapping threshold value as a result,
The street view image with path adaptation of not a large amount of repeated acquisitions accurately shows the streetscape figure to match with road in panoramic table
Picture improves efficiency when street view image and path adaptation.
Based on above-mentioned optimization, as shown in Figure 2 a, technical solution provided in this embodiment is specific as follows:
S210: the image for determining target road and matching with target road.
S220: according to the position of the collection point of each described image, projection position of the collection point in the target road is determined
It sets.
In the present embodiment, it is illustrated, as shown in Figure 2 b, straight line L characterizes target road, matches with target road
Image there are 8 images, and the collection point of 8 images is respectively a, b, c, d, e, f, g and h, due to the distribution of 8 collection points
Not on the same line, therefore 8 collection points are projected on target road L respectively, then the projection of 8 collection points
Position is respectively the position where a ', b ', c ', d ', e ', f ', g ' and h '.
S230: the image that the distance of projected position is less than setting overlapping threshold value is subjected to duplicate removal processing.
In the present embodiment, setting overlapping threshold value is set as needed.It is set specifically, the distance of projected position is less than
Surely be overlapped threshold value image carry out duplicate removal processing after, guarantee adjacent projected position the distance between be equal to or more than
Setting overlapping threshold value.As shown in Figure 2 b, the distance between c ' and f ' are equal to setting overlapping threshold value, and projected position is in c '
Image between f ' is deleted, i.e., by projected position respectively in d ' and e ' two images delete.Figure after deletion
The collection point of picture is as shown in Figure 2 c, as shown, since the distance between adjacent projected position is greater than or equal to setting overlapping
Threshold value, therefore there is no the duplicate images with path adaptation in image in the database.
The recognition methods of a kind of street view image provided in this embodiment, by the way that the distance of projected position is less than setting overlapping
The image of threshold value carries out duplicate removal processing, can identify it is a large amount of repeated acquisitions with path adaptation street view image, in panorama
Figure accurately shows the street view image to match with road, improves efficiency when street view image and path adaptation.
Embodiment three
Fig. 3 a is a kind of recognition methods flow chart of street view image provided in an embodiment of the present invention, in above-described embodiment one
On the basis of, optionally, according to the position of the collection point of each described image, determine projection position of the collection point in the target road
It sets, and screening is carried out to each described image according to projected position and includes:
Acquisition time from each described image according to collecting vehicle is divided at least two image collections, and according to it is described extremely
The weighted value of few two image collections determines target image set；
The collection point position of the image in each image collection is projected in the target road, to determine projected position；
By the projected position of image each in non-object image set, one by one with the image projection positions in target image set
It is compared；
If the distance of projected position is less than setting overlapping threshold value, by the image-erasing in non-object image set；
If the distance of projected position is greater than or equal to setting overlapping threshold value, by the image in non-object image set
It is inserted into the target image set according to projected position.
As a result, by the way that at least two image collections will be divided into the image that target road matches, and choose target figure
Image set close, by the projected position of image each in non-object image set, one by one with the image projection positions in target image set
It is compared；The image in non-targeted combination is inserted into the target image set according to projected position based on comparative result
In, it can according to need the image retained in target image set, and based on the image version in target image set, avoid
The problem of image occurred during different images versions and path adaptation jumps.
Based on above-mentioned optimization, as shown in Figure 3a, technical solution provided in this embodiment is specific as follows:
S310: the image for determining target road and matching with target road.
S320: the acquisition time from each described image according to collecting vehicle is divided at least two image collections, according to institute
The weighted value for stating at least two image collections determines target image set.
It in the present embodiment, can be using the maximum image collection of weight selection value as mesh when determining target image set
The weighted value of logo image set, image collection can be by the acquisition time of image, the position of the collection point of image or other factors
It is determined.For example, if user needs with the image version of newest acquisition for main version, by last date collection
Image collection is as target image set.
S330: the collection point position of the image in each image collection is projected in the target road, to determine projection
Position.
In the present embodiment, for example, as shown in Figure 3b, straight line L characterizes target road, with the matched figure of target road
Picture is there are 12, and 12 Image Acquisition points are respectively as follows: a, b, c, d, e, f, g, h, i, j, k and l.Divide an image into target figure
Image set closes and non-object image set, and non-object image set includes that the conjunction of the first image subset and the second image subset are closed.Wherein,
In target image set the collection point position of image be a, b, c and d, and in target image set the collection point of image projection position
It is set to the position where a ', b ', c ' and d '.First image subset close in the collection point position of image be e, f, g and h, and first
The projected position of the collection point of image is the position where e ', f ', g ' and h ' in image subset conjunction.Second image subset is schemed in closing
The collection point position of picture is respectively i, j, k, l, and the second image subset close in image collection point projected position be respectively i ',
Position where j ', k ' and l '.
S340: the projected position of image each in non-object image set is thrown with the image in target image set one by one
Shadow position is compared.
In the present embodiment, for example, as shown in Figure 3b, by the projected position of image each in non-object image set
(position where e ', f ', g ', h ', i ', j ', k ' and l ') one by one in target image set projected position (a ', b ', c ' and
D ') where position be compared.
S350: if the distance of projected position is less than setting overlapping threshold value, by the image in non-object image set
It deletes.
In the present embodiment, the distance of projected position refers to the distance between projected position.As shown in Figure 3b, if e '
The distance between the distance between c ', f ' and d ' be respectively less than the Chong Die threshold value of setting, and projected position is adopted in e ', f ' respectively
Collection point (e and f) is deleted, and the image that collection point is respectively e, f is deleted.
S360:, will be in non-object image set if the distance of projected position is greater than or equal to setting overlapping threshold value
Image be inserted into the target image set according to projected position.
In the present embodiment, if the distance between g ' and d ' are greater than or equal to setting overlapping threshold value, position will be projected
The image set in g ' is inserted into target image set, and is retained in the collection point g ' projected position.If i ' and g ' it
Between distance be less than setting overlapping threshold value, by projected position i ' collection point i delete, and by collection point be i image into
Row is deleted.If the distance between g ' and h ' are greater than or equal to setting overlapping threshold value, the image by projected position in h ' is inserted
Enter into target image set, and projected position is retained in the collection point h '.It is set if the distance between h ' and j ' are less than
Surely it is overlapped threshold value, the collection point j by projected position in j ' is deleted, and the image that collection point is j is deleted.By above-mentioned
The method for being inserted into target image set, projected position is also inserted into target image set in the image of k ', l '.Pass through
The position of above-mentioned method, the Image Acquisition point of formation is as shown in Figure 3c.By above-mentioned method, it can identify that a large amount of repetitions are adopted
The image of collection, and can be needed to retain the image in target image set according to user, avoid different images version and road
The problem of image jump occurred in the matching process of road.
On the basis of the above embodiments, described when including that at least two image subsets are closed in non-object image set
By the projected position of image each in non-object image set, it is compared one by one with the image projection positions in target image set
Include: the weighted value closed according at least two image subset, determines the throwing of image in at least two image subsets conjunction
The comparative sequence of image projection positions in shadow position and target image set.
For example, as shown in Figure 3b, if the weighted value that the second image subset is closed is greater than the weighted value that the first image subset is closed.
The projected position of image in the conjunction of second image subset is compared with image projection positions in target image set.Such as Fig. 3 b institute
Show, if the distance between d ' and i ' are greater than or equal to the Chong Die threshold value of setting, the image by projected position in i ' is inserted into
In target image set, and the collection point i to projected position in i ' retains.If the distance between j ' and i ' are greater than or wait
It is overlapped threshold value in setting, then is inserted into projected position in target image set in the image of j ', and to projected position in j '
Collection point retained.Similarly, projected position is inserted into target image set in the image of k ', l ' respectively, and to throwing
Shadow position is retained in the collection point of k ', l ' respectively.The image in the conjunction of the second image subset all retains as a result, works as target
In image collection in the conjunction of image projection positions and the second image subset compared with the projected position of image after, then with the first image
The projected position of image is compared in subclass.As shown in Figure 3b, if between the distance between e ' and c ', f ' and d ' away from
From the distance between, g ' and i ', the distance between h ' and i ', respectively less than setting overlapping door restrict, by projected position respectively e ',
The collection point (e, f, g and h) of f ', g ' and h ' are deleted, and the image that collection point is respectively e, f, g, h is deleted.Shape
At Image Acquisition point position it is as shown in Figure 3d.As shown in figs. 3 c and 3d, since the first image subset is closed and the second image
The projected position of image is different from the projected position comparative sequence of image in target image set in set, therefore the image retained is adopted
Collection point is different, and the image of reservation is also different.Therefore, by above-mentioned method, non-targeted figure can be retained according to the needs of users
Image version in image set conjunction.
On the basis of the above embodiments, when including that at least two image subsets are closed in non-object image set, in order to subtract
Mistake when few target image set is compared with image projection positions in non-object image set, according to adopting for each described image
The position for collecting point, determines projected position of the collection point in the target road, and according to projected position to each described image into
Row screening can also be following operating procedure: be divided at least two according to the acquisition time of collecting vehicle from each described image
Image collection, and target image set is determined according to the weighted value of at least two image collection；In non-object image set
According to weighted value size selection be inserted into image subset close；By target image set and be inserted into image subset close in scheme
The collection point position of picture, is projected in the target road, to determine projected position；Each image in image collection will be inserted into
Projected position is compared with the image projection positions in target image set one by one；It is set if the distance of projected position is less than
Surely it is overlapped threshold value, then will be inserted into the image-erasing in image subset conjunction；It is set if the distance of projected position is greater than or equal to
Surely it is overlapped threshold value, then is inserted into the image being inserted into image subset conjunction in the target image set according to projected position；
The operation for executing and choosing and being inserted into image subset conjunction is returned to, until completing the insertion process to the non-object image set.
On the basis of the above embodiments, further includes: if there are the projections of at least one image in target image set
Position is located at intersection, and the image that projected position in non-object image set is located at the intersection is deleted.Its
In, according to data such as the length of the road stored in database, width, position etc. can the intersection to every road carry out
Judgement.It is matched in this way, can retain according to the needs of users in target image set with the intersection of target road
Image.
A kind of recognition methods of street view image provided in this embodiment is divided by the image that will be matched with target road
For at least two image collections, and target image set is chosen, by the projected position of image each in non-object image set, one by one
It is compared with the image projection positions in target image set；Based on comparative result by the image in non-targeted combination according to throwing
Shadow position is inserted into the target image set, can according to need the image retained in target image set, and with target
Based on image version in image collection, showing for the image jump occurred during different images version and path adaptation is avoided
As.
Example IV
Fig. 4 is a kind of identification device structural block diagram for street view image that the embodiment of the present invention four provides, and described device is used for
Execute the recognition methods of street view image.As shown in figure 4, described device comprises determining that module 410 and screening module 420.
Wherein it is determined that module 410, the image for determining target road and matching with target road；
Screening module 420 determines collection point in the target road for the position according to the collection point of each described image
On projected position, and each described image is screened according to projected position.
Further, the screening module 420, is specifically used for:
According to the position of the collection point of each described image, projected position of the collection point in the target road is determined；
The image that the distance of projected position is less than setting overlapping threshold value is subjected to duplicate removal processing.
Further, the screening module 420, is specifically used for:
Acquisition time from each described image according to collecting vehicle is divided at least two image collections, and according to it is described extremely
The weighted value of few two image collections determines target image set；
The collection point position of the image in each image collection is projected in the target road, to determine projected position；
By the projected position of image each in non-object image set, one by one with the image projection positions in target image set
It is compared；
If the distance of projected position is less than setting overlapping threshold value, by the image-erasing in non-object image set；
If the distance of projected position is greater than or equal to setting overlapping threshold value, by the image in non-object image set
It is inserted into the target image set according to projected position.
Further, the non-object image set is closed including at least two image subsets；
The projected position by image each in non-object image set, one by one with the image projection in target image set
Position is compared, comprising:
According to the weighted value that at least two image subset is closed, image in at least two image subsets conjunction is determined
The comparative sequence of image projection positions in projected position and target image set.
Further, described device further include: intersection processing module 430, if for being deposited in target image set
It is located at intersection in the projected position of at least one image, projected position in non-object image set is located at the crossroad
The image of mouth is deleted.
Method provided by any embodiment of the invention can be performed in above-mentioned apparatus, has the corresponding functional module of execution method
And beneficial effect.
Technical solution provided in an embodiment of the present invention, by the determining Image Acquisition point to match with target road in target
Projected position on road, and screened according to projected position pair and the image that target road matches, it can identify a large amount of
The street view image with path adaptation of repeated acquisition accurately shows the street view image to match with road in panoramic table, mentions
Efficiency when high street view image and path adaptation.
Note that the above is only a better embodiment of the present invention and the applied technical principle.It will be appreciated by those skilled in the art that
The invention is not limited to the specific embodiments described herein, be able to carry out for a person skilled in the art it is various it is apparent variation,
It readjusts and substitutes without departing from protection scope of the present invention.Therefore, although being carried out by above embodiments to the present invention
It is described in further detail, but the present invention is not limited to the above embodiments only, without departing from the inventive concept, also
It may include more other equivalent embodiments, and the scope of the invention is determined by the scope of the appended claims.
Claims (8)
1. a kind of recognition methods of street view image characterized by comprising
The image for determining target road and matching with target road；
According to the position of the collection point of each described image, projected position of the collection point in the target road is determined, and according to
Projected position screens each described image；
Wherein, according to the position of the collection point of each described image, projected position of the collection point in the target road is determined, and
Each described image is screened according to projected position, comprising:
At least two image collections will be divided into the image that target road matches, and choose target image set, by non-mesh
The projected position of each image in logo image set, is compared with the image projection positions in target image set one by one, is based on
Image in non-targeted set is inserted into the target image set by comparison result according to projected position.
2. the method according to claim 1, wherein by being divided at least two with the image that target road matches
A image collection, and choose target image set, by the projected position of image each in non-object image set, one by one with target figure
Image projection positions in image set conjunction are compared, based on comparative result by the image in non-targeted set according to projected position
It is inserted into the target image set, comprising:
Acquisition time from each described image according to collecting vehicle is divided at least two image collections, and according to described at least two
The weighted value of a image collection determines target image set；
The collection point position of the image in each image collection is projected in the target road, to determine projected position；
By the projected position of image each in non-object image set, one by one with the image projection positions in the target image set
It is compared；
If the distance of projected position is less than setting overlapping threshold value, by the image-erasing in the non-object image set；
If the distance of projected position is greater than or equal to setting overlapping threshold value, by the image in the non-object image set
It is inserted into the target image set according to projected position.
3. according to the method described in claim 2, it is characterized in that, the non-object image set includes at least two images
Set；
The projected position by image each in non-object image set, one by one with the image projection in the target image set
Position is compared, comprising:
According to the weighted value that at least two image subset is closed, the projection of image in at least two image subsets conjunction is determined
The comparative sequence of image projection positions in position and the target image set.
4. according to the method described in claim 2, it is characterized by further comprising:
It, will be described non-targeted if there are the projected positions of at least one image to be located at intersection in the target image set
The image that projected position is located at the intersection in image collection is deleted.
5. a kind of identification device of street view image characterized by comprising
Determining module, the image for determining target road and matching with target road；
Screening module determines throwing of the collection point in the target road for the position according to the collection point of each described image
Shadow position, and each described image is screened according to projected position；
Wherein, the screening module is specifically used for:
At least two image collections will be divided into the image that target road matches, and choose target image set, by non-mesh
The projected position of each image in logo image set, is compared with the image projection positions in target image set one by one, is based on
Image in non-targeted set is inserted into the target image set by comparison result according to projected position.
6. device according to claim 5, which is characterized in that the screening module is specifically used for:
Acquisition time from each described image according to collecting vehicle is divided at least two image collections, and according to described at least two
The weighted value of a image collection determines target image set；
The collection point position of the image in each image collection is projected in the target road, to determine projected position；
By the projected position of image each in non-object image set, one by one with the image projection positions in the target image set
It is compared；
If the distance of projected position is less than setting overlapping threshold value, by the image-erasing in the non-object image set；
If the distance of projected position is greater than or equal to setting overlapping threshold value, by the image in the non-object image set
It is inserted into the target image set according to projected position.
7. device according to claim 6, which is characterized in that the non-object image set includes at least two images
Set；
The projected position by image each in non-object image set, one by one with the image projection in the target image set
Position is compared, comprising:
According to the weighted value that at least two image subset is closed, the projection of image in at least two image subsets conjunction is determined
The comparative sequence of image projection positions in position and the target image set.
8. device according to claim 6, which is characterized in that further include intersection processing module, if for described
There are the projected positions of at least one image to be located at intersection in target image set, will throw in the non-object image set
The image that shadow position is located at the intersection is deleted.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610865626.8A CN106372255B (en) | 2016-09-29 | 2016-09-29 | A kind of recognition methods of street view image and device |
| PCT/CN2017/074300 WO2018058888A1 (en) | 2016-09-29 | 2017-02-21 | Street view image recognition method and apparatus, server and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610865626.8A CN106372255B (en) | 2016-09-29 | 2016-09-29 | A kind of recognition methods of street view image and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106372255A CN106372255A (en) | 2017-02-01 |
| CN106372255B true CN106372255B (en) | 2019-01-22 |
Family
ID=57898310
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610865626.8A Active CN106372255B (en) | 2016-09-29 | 2016-09-29 | A kind of recognition methods of street view image and device |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN106372255B (en) |
| WO (1) | WO2018058888A1 (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106372255B (en) * | 2016-09-29 | 2019-01-22 | 北京百度网讯科技有限公司 | A kind of recognition methods of street view image and device |
| CN108460815B (en) * | 2017-02-22 | 2022-06-17 | 腾讯科技（深圳）有限公司 | Method and device for editing map road elements |
| CN110099237B (en) * | 2018-01-31 | 2021-08-17 | 腾讯科技（深圳）有限公司 | Image processing method, electronic device, and computer-readable storage medium |
| CN109902134B (en) * | 2019-02-18 | 2021-10-22 | 北京市天元网络技术股份有限公司 | Method and device for accurately presenting fault road section in engineering road and electronic equipment |
| CN110795994B (en) * | 2019-09-16 | 2022-05-10 | 腾讯科技（深圳）有限公司 | Intersection image selection method and device |
| CN110853098B (en) * | 2019-11-15 | 2023-06-16 | 上海思岚科技有限公司 | Robot positioning method, device, equipment and storage medium |
| CN117216301A (en) * | 2023-11-08 | 2023-12-12 | 高德软件有限公司 | Image data recommendation method and device, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103914521A (en) * | 2014-03-20 | 2014-07-09 | 北京建筑大学 | Street view image storage method and device based on mixed tile pyramids |
| CN104268513A (en) * | 2014-09-17 | 2015-01-07 | 百度在线网络技术（北京）有限公司 | Method and device for acquiring road guide data |
| CN105205837A (en) * | 2014-06-25 | 2015-12-30 | 上海照炎信息科技有限公司 | Position aggregation-based three-dimensional panoramic picture sparsification method |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9215448B2 (en) * | 2013-01-31 | 2015-12-15 | Here Global B.V. | Stereo panoramic images |
| US9986154B2 (en) * | 2014-05-21 | 2018-05-29 | Here Global B.V. | Developing a panoramic image |
| CN106372255B (en) * | 2016-09-29 | 2019-01-22 | 北京百度网讯科技有限公司 | A kind of recognition methods of street view image and device |
-
2016
- 2016-09-29 CN CN201610865626.8A patent/CN106372255B/en active Active
-
2017
- 2017-02-21 WO PCT/CN2017/074300 patent/WO2018058888A1/en active Application Filing
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103914521A (en) * | 2014-03-20 | 2014-07-09 | 北京建筑大学 | Street view image storage method and device based on mixed tile pyramids |
| CN105205837A (en) * | 2014-06-25 | 2015-12-30 | 上海照炎信息科技有限公司 | Position aggregation-based three-dimensional panoramic picture sparsification method |
| CN104268513A (en) * | 2014-09-17 | 2015-01-07 | 百度在线网络技术（北京）有限公司 | Method and device for acquiring road guide data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106372255A (en) | 2017-02-01 |
| WO2018058888A1 (en) | 2018-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106372255B (en) | A kind of recognition methods of street view image and device | |
| CN109741257B (en) | Full-automatic panorama shooting and splicing system and method | |
| CN106570446B (en) | The method and apparatus of lane line drawing | |
| CN104866500B (en) | Picture classification display method and device | |
| JP6175846B2 (en) | Vehicle tracking program, server device, and vehicle tracking method | |
| CN110062871A (en) | Method and system for positioning and mapping based on video | |
| Azimi et al. | Eagle: Large-scale vehicle detection dataset in real-world scenarios using aerial imagery | |
| CN102804231A (en) | Piecewise planar reconstruction of three-dimensional scenes | |
| US20220398856A1 (en) | Method for reconstruction of a feature in an environmental scene of a road | |
| CN110348143A (en) | A kind of track merging method, device and storage medium | |
| CN108241819A (en) | The recognition methods of pavement markers and device | |
| WO2021154155A1 (en) | Methods and data processing systems for predicting road attributes | |
| JP4131930B2 (en) | Map update program and map update device | |
| CN109102026A (en) | A kind of vehicle image detection method, apparatus and system | |
| CN112149471B (en) | Loop detection method and device based on semantic point cloud | |
| Champion et al. | A test of automatic building change detection approaches | |
| CN110874170A (en) | Image area correction method, image segmentation method and device | |
| CN110827340B (en) | Map updating method, device and storage medium | |
| CN111738061A (en) | Binocular vision stereo matching method based on regional feature extraction and storage medium | |
| CN107808160B (en) | Three-dimensional building extraction method and device | |
| CN109598199A (en) | Lane line generation method and device | |
| CN112507887B (en) | Intersection sign extracting and associating method and device | |
| JP4963824B2 (en) | Change area extraction apparatus and change area extraction method | |
| JP6384050B2 (en) | Line detection apparatus, line detection method, and program | |
| Vasudha et al. | Carriageway Edge Detection for Unmarked Urban Roads using Deep Learning Techniques |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |