CN105957005A - Method for bridge image splicing based on feature points and structure lines - Google Patents
Method for bridge image splicing based on feature points and structure lines Download PDFInfo
- Publication number
- CN105957005A CN105957005A CN201610269079.7A CN201610269079A CN105957005A CN 105957005 A CN105957005 A CN 105957005A CN 201610269079 A CN201610269079 A CN 201610269079A CN 105957005 A CN105957005 A CN 105957005A
- Authority
- CN
- China
- Prior art keywords
- image
- air strips
- point
- panorama sketch
- bridge
- 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
- 238000000034 method Methods 0.000 title claims abstract description 63
- 238000012545 processing Methods 0.000 claims abstract description 7
- 238000013480 data collection Methods 0.000 claims description 36
- 238000000605 extraction Methods 0.000 claims description 27
- 230000008878 coupling Effects 0.000 claims description 13
- 238000010168 coupling process Methods 0.000 claims description 13
- 238000005859 coupling reaction Methods 0.000 claims description 13
- 238000000265 homogenisation Methods 0.000 claims description 8
- 239000000284 extract Substances 0.000 claims description 4
- 230000015572 biosynthetic process Effects 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 239000000203 mixture Substances 0.000 claims description 2
- 238000011524 similarity measure Methods 0.000 claims description 2
- 230000000694 effects Effects 0.000 abstract description 6
- 238000007796 conventional method Methods 0.000 abstract 1
- 230000001419 dependent effect Effects 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 4
- 238000005286 illumination Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 230000001186 cumulative effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000008030 elimination Effects 0.000 description 2
- 238000003379 elimination reaction Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 230000001629 suppression Effects 0.000 description 2
- 206010028980 Neoplasm Diseases 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 239000010931 gold Substances 0.000 description 1
- 229910052737 gold Inorganic materials 0.000 description 1
- 230000009916 joint effect Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/14—Transformations for image registration, e.g. adjusting or mapping for alignment of images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/32—Indexing scheme for image data processing or generation, in general involving image mosaicing
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention provides a method for bridge image splicing based on feature points and structure lines. The method comprises the following steps: preparing data; dividing a strip data set; projecting a strip panorama; projecting a panorama of a bridge bottom; processing the image; and outputting the result. The method provided by the invention solves the technical problem that a conventional method in the prior art is excessively dependent on the initial condition or the image texture splicing effect.
Description
Technical field
The present invention relates to photogrammetric and computer vision field, be specifically related to a kind of feature based
Point and the bridge image splicing method of structure lines.
Background technology
Image splicing method is Digital Image Processing, the photogrammetric and weight of computer vision field
Wanting one of research topic, it is widely used in geographical national conditions monitoring, medical science tumor examination and mixes
Solidifying civil engineering builds Defect inspection etc..Bottom concrete-bridge in health detection, in order to obtain high score
Image bottom the bridge of resolution, intelligent CCD camera can only cover the least bottom bridge one
Point, therefore, it is i.e. for a small bridge, also tends to need to shoot thousands of up to a hundred figures
Can be good at obtaining whole bridge footer information.In the face of the data of such magnanimity, the most efficiently
A complete bridge bottom surface panorama sketch that these image joints are become just become bridge health inspection
Technology the most key in looking into.
Image splicing method mainly has two kinds of technology paths at present, and one is direct image splicing method,
Another kind is the image splicing method of feature based.
Directly image splicing method is to utilize all of image data directly to do image joint, and this is past
Toward obtaining the image joint of degree of precision, it require that certain initial value or preferably
Geometric correction.The major defect of the method is the dependency for initial condition, splices result
Quality be largely depending on the precision of image initial value.
The image splicing method of feature based is that the characteristic information utilizing image is to carry out image spelling
The method connect, this method compares faster and robust, it is possible to well registration has overlapping region
Image set.The texture of the limitation of the method mainly image can not be guaranteed, image
The effect of texture-rich degree image joint.
Summary of the invention
Initial condition or image texture is excessively relied on for solving the image splicing method of prior art
The technical problem of image joint effect, the present invention provides one to be independent of initial condition, and splicing
The bridge image splicing method of effective distinguished point based and structure lines.
The bridge image splicing method of a kind of distinguished point based and structure lines, specifically includes following step
Rapid:
Step 1: data prepare: prepare data acquisition Route Planning Data collection H, image data
Collection E, image initial attitude data set P and bridge 3D cloud data collection M, and set up 3D bridge
Beam model；
Step 2: divide air strips data set: according to described image initial attitude data set P by institute
State image data collection E and be divided into multiple air strips data set S1,S2,S3,…,SN, thus constitute image
CollectionDescribed image set
Step 3: projection air strips panorama sketch: for described image setIn arbitrary described air strips number
According to collection Sk, k=1,2 ... N, the image data in the data set of described air strips is registrated and throws
Shadow becomes an air strips panorama sketch Wk, k=1,2 ... N, thus constitute image set η, described image
Collection
Step 4: panorama sketch bottom projection bridge: by the whole described boat in described image set η
Band panorama sketch registrates, and projects into panorama sketch O bottom a complete bridge；
Step 5: image processing: panorama sketch O bottom described bridge is carried out splicing line lookup,
And carry out light and color homogenization process；
Step 6: output result.
The present invention provide distinguished point based and structure lines bridge image splicing method one
In kind of preferred embodiment, described step 2 specifically includes following steps:
Step 21: obtain image attitude key point set: according to described data acquisition path planning
Data set H, obtains image attitude key point set during data acquisition；
Step 22: image initial attitude data set P is grouped: according to described image attitude key point
Position in collection and attitude data, be grouped, often described image initial attitude data set P
Image appearances different in described image attitude key point set during individual packet corresponding data collection respectively
State key point；
Step 23: divide air strips data set: according to initial to described image in described step 22
The packet that attitude data collection P is carried out, is divided into multiple described air strips number by described image data collection E
According to collection S1,S2,S3,…,SN, each described air strips data set corresponding initial appearance of described image respectively
One packet of state data set P, thus constitute described image setDescribed image set
The present invention provide distinguished point based and structure lines bridge image splicing method one
In kind of preferred embodiment, described step 3 specifically includes following steps:
Step 31: divide image data collection: by each described air strips data set Sk, k=
1,2 ... N is further divided into multiple image data collection G1,G2,G3,…,GB, each described image
Data set Gk, k=1,2 ... B comprises the image of moderate number respectively, thus constitutes image set δ,
Described image set
Step 32: projection packet panorama sketch: for each described image in described image set δ
Data set Gk, k=1,2 ... B, carry out respectively registrating and projecting composition by its interior image data
Group panorama sketch Uk, k=1,2 ... B, thus constitute image set τ, described image set
Step 33: project individual air strips panorama sketch: by complete for the described packet in described image set τ
Scape figure UkRegistrate, project into individual described air strips panorama sketch Wk；
Step 34: project complete air strips panorama sketch: repeating said steps 31 arrives described step 33,
By described image setIn each described air strips data set SkRegistrate respectively, project into complete
Described air strips panorama sketch Wk, k=1,2 ... N, thus constitute described image set η, described image
Collection
The present invention provide distinguished point based and structure lines bridge image splicing method one
In kind of preferred embodiment, described step 32 specifically includes following steps:
Step 321: feature point extraction and extraction of structure lines: to each described image data collection Gk,
K=1,2 ... B carries out feature point extraction and extraction of structure lines respectively, and wherein feature point extraction selects
SIFT feature point and ShiTomasi characteristic point；Extraction of structure lines then utilizes described bridge 3D point
The 3D structure lines that cloud data set M extracts, by the demarcation between camera, finds 3D to tie
2D structure lines on image corresponding to structure line；
Step 322: pyramid image mates: set up image pyramid, utilizes at described image
The SIFT feature point that pyramid is high-rise, the list that coupling is set up between image should be related to；By institute
State the list between image should be related to, on low layer pyramid, utilize ShiTomasi feature to click on
Row similarity measure mates, thus obtains precision preferable Image Matching point；
Described estimate matching formula and be expressed as:
Wherein,WithIt it is the m × n window chosen centered by characteristic pointWithThe meansigma methods of middle pixel, its computing formula is as follows:
Wherein, I (q) and I ' (q ') is the m × n window chosen centered by characteristic pointWithThe value of middle pixel, ρ (p, p ') is closer to 1, then the similarity between pixel is the highest,
Mate the most reliable；
The described Image Matching point obtained by Image Matching, thus set up described image data collection
GkCorresponding relation between interior image and image；
Step 323: obtain optimum image attitude: described in obtaining in described step 321
The described Image Matching point obtained in 2D structure lines and described step 322, more smart to obtain
True image attitude；
Energy-optimised function includes El(θ) and Ep(θ) two:
E (θ)=arg min (El(θ)+Ep(θ)),
Wherein El(θ) and Ep(θ) energy term of described 2D structure lines and described image are represented respectively
The energy term of match point,
Energy term E of described 2D structure linesl(θ) be calculated as follows,
Wherein, p is a series of value point such as grade on the described 2D structure lines of this image,Generation
Table is the attitude parameter of this image,Wait value spot projection to described 3D bridge for described
3D point on beam model,For distance on described 3D structure linesNearest 3D
Point；
Energy term E of described Image Matching pointp(θ) be calculated as follows,
Wherein, what I represented is coupling image, and what x and x ' represented is the match point on coupling image
Collection,WithRepresent is that the described Image Matching spot projection on coupling image is to institute
State spatial point corresponding on 3D bridge model；
By using LM function optimization, the initial attitude of image is adjusted, until function
Optimize energy term minimum, thus the optimum image attitude obtained；
Step 324: project individual packet panorama sketch: utilize the image attitude after optimizing by described
Image data collection GkProject on described 3D bridge model, obtain a described packet panorama sketch
Uk。
Step 325: projection is all grouped panorama sketch: repeating said steps 321 arrives described step
324, each packet in described image set τ is registrated respectively, and projects into complete described packet
Panorama sketch Uk, k=1,2 ... B, thus constitute described image set τ, described image set
The present invention provide distinguished point based and structure lines bridge image splicing method one
In kind of preferred embodiment, described step 33 specifically includes following steps:
Step 331: setting up packet list should be related to: utilize adjacent described packet panorama sketch UkIt
Between the common region that has, set up described packet panorama sketch UkBetween list should be related to, and lead to
Cross described packet panorama sketch UkBetween list should be related to, by adjacent described packet panorama sketch Uk
Merge；
Step 332: project individual air strips panorama sketch: utilize the institute that described step 331 is set up
State packet panorama sketch UkBetween list should be related to, by each described packet panorama of described image set τ
Figure UkMerge one by one, the described air strips panorama sketch W of final acquisitionk。
The present invention provide distinguished point based and structure lines bridge image splicing method one
In kind of preferred embodiment, described step 4 specifically includes following steps:
Step 41: set up air strips corresponding relation: by described air strips data set SkEach interior image,
Respectively with adjacent described air strips data set SkThe image inside with degree of overlapping carries out correspondence, thus
Set up adjacent air strips corresponding relation；
Step 42: set up air strips and singly should be related to: utilize described adjacent air strips corresponding relation, build
Vertical adjacent air strips singly should be related to；
Step 43: project adjacent air strips panorama sketch: utilize described adjacent air strips singly should be related to,
By panorama sketch W in air strips described inkProject to another adjacent described air strips panorama sketch WkOn；
Step 44: panorama sketch bottom projection bridge: repeating said steps 41 to described step 43,
Constantly by panorama sketch W in air strips described inkProject to another adjacent described air strips panorama sketch Wk
On, thus set up panorama sketch O bottom complete described bridge.
The present invention provide distinguished point based and structure lines bridge image splicing method one
In kind of preferred embodiment, described step 41 specifically includes following steps:
Step 411: scaling: to described air strips data set SkInterior each image zooms in and out, and carries
Take its characteristic point, and carry out feature description, it is thus achieved that feature descriptor；
Step 412: characteristic matching: by described air strips data set SkInterior each image, respectively
With adjacent described air strips data set SkAll of image carries out characteristic matching one by one, i.e. calculates institute
Stating the Euclidean distance between feature descriptor, the point of selected distance minimum is as optimal match point；
Step 413: set up air strips corresponding relation: the number according to described optimal match point is how many
Select optimal image to match, thus set up described adjacent air strips corresponding relation.
The present invention provide distinguished point based and structure lines bridge image splicing method one
In kind of preferred embodiment, described step 42 specifically includes following steps:
Step 421: air strips feature point extraction: utilize described adjacent air strips data set corresponding relation,
Carrying out feature point extraction wherein on a pair image respectively, feature point extraction uses SIFT
Characteristic point and ShiTomasi characteristic point；
Step 422: air strips pyramid image mates: on the image of described step 421, weight
The operation of multiple described step 322, thus obtain precision preferable air strips data set Image Matching point；
Step 423: set up air strips and singly should be related to: utilize described air strips data set Image Matching point,
Set up described adjacent air strips by matching double points singly should be related to.
The present invention provide distinguished point based and structure lines bridge image splicing method one
In kind of preferred embodiment, described step 5 specifically includes following steps:
Step 51: image set is grouped: be grouped all image set being registered,
Moderate image number is comprised in making often to organize；
Step 52: image joint: to the image in each packet, carries out splicing line simultaneously and looks into
Look for and processing with light and color homogenization, form a big figure；
Step 53: packet splicing: the big figure to the formation of each packet, splices again
Line is searched and light and color homogenization processes, and ultimately forms image panorama sketch at the bottom of a complete bridge.
Compared to prior art, the described distinguished point based of present invention offer and the bridge of structure lines
Image splicing method has the advantages that
One, the bridge image splicing method of described distinguished point based and structure lines is well by thousands of
What bottom the bridge under upper different illumination conditions up to a hundred, image was complete be spliced into one has identical
The bridge bottom surface panorama sketch of texture color, has the highest robustness, accuracy, either spells
Connect local error or global error, all obtained effective suppression, reach preferable splicing effect
Really.
Two, the bridge image splicing method of described distinguished point based and structure lines not only makes full use of
The texture information of image and half-tone information, have also combined the structural information of bridge itself so that
The drift error of Image Matching has obtained good elimination.Situation about either lacking for texture
Or the situation that illumination condition is different, all obtains good effect, has been sufficiently reproduced bottom bridge
Situation, it is simple to carry out bridge health location and details observe.
Three, the bridge image splicing method of described distinguished point based and structure lines introduces structure lines work
For the energy term of image joint, well inhibit the drift error of image.Compare traditional only
Utilizing the image joint that characteristic point is done, the most well eliminate during image joint is complete
Office's cumulative error, in hgher efficiency, better, it is adaptable to the image of multiple different bridge types
Splicing, well solves the image joint problem of big data quantity.
Accompanying drawing explanation
For the technical scheme being illustrated more clearly that in the embodiment of the present invention, below will be to embodiment
Accompanying drawing used in description is briefly described, it should be apparent that, the accompanying drawing in describing below
It is only some embodiments of the present invention, for those of ordinary skill in the art, is not paying
On the premise of going out creative work, it is also possible to obtain other accompanying drawing according to these accompanying drawings, its
In:
Fig. 1 is the distinguished point based that provides of the present invention and the bridge image splicing method of structure lines
Flow chart；
Fig. 2 is the distinguished point based that provides of the present invention and the bridge image splicing method of structure lines
Structure lines simulation schematic diagram；
Fig. 3 is the distinguished point based that provides of the present invention and the bridge image splicing method of structure lines
Image Matching schematic flow sheet；
Fig. 4 is the distinguished point based that provides of the present invention and the bridge image splicing method of structure lines
Image pose refinement schematic diagram.
Detailed description of the invention
Below in conjunction with the accompanying drawing in the embodiment of the present invention, to the technical side in the embodiment of the present invention
Case is clearly and completely described, it is clear that described embodiment is only of the present invention
Divide embodiment rather than whole embodiments.
Refer to Fig. 1, be that the bridge image of the distinguished point based that provides of the present invention and structure lines is spelled
Connect the flow chart of method.
The bridge image splicing method 1 of described distinguished point based and structure lines specifically includes following
Step:
S1: data prepare: prepare data acquisition Route Planning Data collection H, image data collection E,
Image initial attitude data set P and bridge 3D cloud data collection M, and set up 3D bridge model.
S2: described S2 specifically includes following steps:
S21: obtain image attitude key point set: according to described data acquisition Route Planning Data
Collection H, obtains image attitude key point set during data acquisition.
S22: image initial attitude data set P packet: according in described image attitude key point set
Position and attitude data, described image initial attitude data set P is grouped, Mei Gefen
During group corresponding data collection respectively, in described image attitude key point set, different image attitudes is closed
Key point.
S23: divide air strips data sets: according in described step 22 to described image initial attitude
The packet that data set P is carried out, is divided into multiple described air strips data set by described image data collection E
S1,S2,S3,…,SN, each described air strips data set corresponding described image initial attitude number respectively
According to a packet of collection P, thus constitute described image setDescribed image set
S3: described S3 specifically includes following steps:
S31: division image data collection: described air strips image set SkComprise image A altogether to open, respectively
For image Ik1,Ik2,Ik3,…,IkA, image quantity is the hugest, processes together and easily expends in a large number
Time and resource, for these images of registration more rapid, efficient, by each described air strips number
According to collection Sk, k=1,2 ... N is further divided into multiple image data collection G1,G2,G3,…,GB,
Each described image data collection Gk, k=1,2 ... B comprises the image of moderate number respectively, thus
Constitute image set δ, described image set
S32: described S32 specifically includes following steps:
Please refer to Fig. 2, it is the distinguished point based that provides of the present invention and the bridge shadow of structure lines
Structure lines simulation schematic diagram as joining method.
S321: feature point extraction and extraction of structure lines: to each described image data collection Gk,
K=1,2 ... B carries out feature point extraction and extraction of structure lines respectively, and wherein feature point extraction selects
SIFT feature point and ShiTomasi characteristic point；Multiple plain splice can be regarded as bottom bridge and
Becoming, and structure lines can be regarded as the intersection of bridge plane and plane, extraction of structure lines utilizes institute
State the 3D structure lines that bridge 3D cloud data collection M extracts, by the demarcation between camera,
Find the 2D structure lines on the image corresponding to 3D structure lines.
Please refer to Fig. 3, it is the distinguished point based that provides of the present invention and the bridge shadow of structure lines
Image Matching schematic flow sheet as joining method.
S322: pyramid image mates: set up image pyramid, utilizes at described image gold word
The SIFT feature point of tower height layer, the list that coupling is set up between image should be related to；By described shadow
List between Xiang should be related to, on low layer pyramid, utilizes ShiTomasi characteristic point to carry out phase
Estimate coupling like property, thus obtain precision preferable Image Matching point.
Described estimate matching formula and be expressed as:
Wherein,WithIt it is the m × n window chosen centered by characteristic pointWithThe meansigma methods of middle pixel, its computing formula is as follows:
Wherein, I (q) and I ' (q ') is the m × n window chosen centered by characteristic pointWithThe value of middle pixel, ρ (p, p ') is closer to 1, then the similarity between pixel is the highest,
Mate the most reliable.
The described Image Matching point obtained by Image Matching, thus set up described image data collection
GkCorresponding relation between interior image and image.
Please refer to Fig. 4, it is the distinguished point based that provides of the present invention and the bridge shadow of structure lines
Image pose refinement schematic diagram as joining method.
S323: obtain optimum image attitude: by the described 2D obtained in described step 321
The described Image Matching point obtained in structure lines and described step 322, more accurate to obtain
Image attitude.
Energy-optimised function includes El(θ) and Ep(θ) two:
E (θ)=arg min (El(θ)+Ep(θ)),
Wherein El(θ) and Ep(θ) energy term of described 2D structure lines and described image are represented respectively
The energy term of match point,
Energy term E of described 2D structure linesl(θ) be calculated as follows,
Wherein, p is a series of value point such as grade on the described 2D structure lines of this image,Generation
Table is the attitude parameter of this image,Wait value spot projection to described 3D bridge for described
3D point on beam model,For distance on described 3D structure linesNearest 3D
Point.
Energy term E of described Image Matching pointp(θ) be calculated as follows,
Wherein, what I represented is coupling image, and what x and x ' represented is the match point on coupling image
Collection,WithRepresent is that the described Image Matching spot projection on coupling image is to institute
State spatial point corresponding on 3D bridge model.
By using LM function optimization, the initial attitude of image is adjusted, until function
Optimize energy term minimum, thus the optimum image attitude obtained.
S324: project individual packet panorama sketch: utilize the image attitude after optimizing by described image
Data set GkProject on described 3D bridge model, obtain a described packet panorama sketch Uk。
S325: projection is all grouped panorama sketch: repeating said steps 321 arrives described step 324,
Each packet in described image set τ is registrated respectively, and projects into complete described packet panorama
Figure Uk, k=1,2 ... B, thus constitute described image set τ, described image set
S33: described S33 specifically includes following steps:
S331: setting up packet list should be related to: utilize adjacent described packet panorama sketch UkBetween
The common region having, sets up described packet panorama sketch UkBetween list should be related to, and pass through
Described packet panorama sketch UkBetween list should be related to, by adjacent described packet panorama sketch UkEnter
Row merges.
S332: project individual air strips panorama sketch: utilize that described step 331 is set up described point
Group panorama sketch UkBetween list should be related to, by each described packet panorama sketch U of described image set τk
Merge one by one, the described air strips panorama sketch W of final acquisitionk。
S34: project complete air strips panorama sketch: repeating said steps 31 arrives described step 33,
By described image setIn each described air strips data set SkRegistrate respectively, project into complete
Described air strips panorama sketch Wk, k=1,2 ... N, thus constitute described image set η, described image
Collection
S4: described S4 specifically includes following steps:
S41: described S41 specifically includes following steps:
S411: scaling: to described air strips data set SkInterior each image zooms in and out, and extracts it
Characteristic point, and carry out feature description, it is thus achieved that feature descriptor.
S412: characteristic matching: by described air strips data set SkInterior each image, respectively with phase
Adjacent described air strips data set SkAll of image carries out characteristic matching one by one, i.e. calculates described spy
Levying the Euclidean distance between descriptor, the point of selected distance minimum is as optimal match point.
S413: set up air strips corresponding relation: how much select according to the number of described optimal match point
Optimal image pairing, thus set up described adjacent air strips corresponding relation.
S42: described S42 specifically includes following steps:
S421: air strips feature point extraction: utilize described adjacent air strips data set corresponding relation,
Carrying out feature point extraction on one pair of which image respectively, feature point extraction uses SIFT special
Levy a little and ShiTomasi characteristic point.
S422: air strips pyramid image mates: on the image of described step 421, repeats institute
State the operation of step 322, thus obtain precision preferable air strips data set Image Matching point.
S423: set up air strips and singly should be related to: utilize described air strips data set Image Matching point is logical
Overmatching point singly should be related to setting up described adjacent air strips.
S43: project adjacent air strips panorama sketch: utilize described adjacent air strips singly should be related to, by one
Described air strips panorama sketch WkProject to another adjacent described air strips panorama sketch WkOn.
S44: panorama sketch bottom projection bridge: repeating said steps 41 to described step 43,
Constantly by panorama sketch W in air strips described inkProject to another adjacent described air strips panorama sketch Wk
On, thus set up panorama sketch O bottom complete described bridge.
S5: described S5 specifically includes following steps:
S51: image set is grouped: is grouped all image set being registered, makes
Often comprise moderate image number in group.
S52: image joint: to the image in each packet, carry out simultaneously splicing line search and
Light and color homogenization processes, and forms a big figure.
S53: packet splicing: the big figure to the formation of each packet, again carries out splicing line and looks into
Look for and processing with light and color homogenization, ultimately form image panorama sketch at the bottom of a complete bridge.
S6: output result.
Compared to prior art, the described distinguished point based of present invention offer and the bridge of structure lines
Image splicing method 1 has the advantages that
One, described distinguished point based well will become with the bridge image splicing method 1 of structure lines
What bottom the bridge under thousand upper different illumination conditions up to a hundred, image was complete be spliced into one has phase
With the bridge bottom surface panorama sketch of texture color, there is the highest robustness, accuracy, either
Splicing local error or global error, all obtained effective suppression, reached preferable splicing
Effect.
Two, the bridge image splicing method 1 of described distinguished point based and structure lines is not only the most sharp
With texture information and the half-tone information of image, have also combined the structural information of bridge itself, make
The drift error obtaining Image Matching has obtained good elimination.The feelings either lacked for texture
Condition or the different situation of illumination condition, all obtain good effect, be sufficiently reproduced at the bottom of bridge
The situation in portion, it is simple to carry out bridge health location and details is observed.
Three, the bridge image splicing method 1 of described distinguished point based and structure lines introduces structure lines
As the energy term of image joint, well inhibit the drift error of image.Compare traditional
The image joint done merely with characteristic point, during the most well eliminating image joint
Overall situation cumulative error, in hgher efficiency, better, it is adaptable to the shadow of multiple different bridge types
As splicing, well solve the image joint problem of big data quantity.
The foregoing is only embodiments of the invention, not thereby limit the scope of the claims of the present invention,
Every equivalent structure utilizing description of the invention content to be made or equivalence flow process conversion, or directly
Or indirectly it is used in other relevant technical field, the most in like manner it is included in the patent protection of the present invention
Within the scope of.
Claims (9)
1. a distinguished point based and the bridge image splicing method of structure lines, it is characterised in that
Specifically include following steps:
Step 1: data prepare: prepare data acquisition Route Planning Data collection H, image data
Collection E, image initial attitude data set P and bridge 3D cloud data collection M, and set up 3D bridge
Beam model；
Step 2: divide air strips data set: according to described image initial attitude data set P by institute
State image data collection E and be divided into multiple air strips data set S1,S2,S3,...,SN, thus constitute image
CollectionDescribed image set
Step 3: projection air strips panorama sketch: for described image setIn arbitrary described air strips number
According to collection Sk, k=1,2 ... N, the image data in the data set of described air strips is registrated and throws
Shadow becomes an air strips panorama sketch Wk, k=1,2 ... N, thus constitute image set η, described image
Collection
Step 4: panorama sketch bottom projection bridge: by the whole described boat in described image set η
Band panorama sketch registrates, and projects into panorama sketch O bottom a complete bridge；
Step 5: image processing: panorama sketch O bottom described bridge is carried out splicing line lookup,
And carry out light and color homogenization process；
Step 6: output result.
The bridge image joint side of distinguished point based the most according to claim 1 and structure lines
Method, it is characterised in that described step 2 specifically includes following steps:
Step 21: obtain image attitude key point set: according to described data acquisition path planning
Data set H, obtains image attitude key point set during data acquisition；
Step 22: image initial attitude data set P is grouped: according to described image attitude key point
Position in collection and attitude data, be grouped, often described image initial attitude data set P
Image appearances different in described image attitude key point set during individual packet corresponding data collection respectively
State key point；
Step 23: divide air strips data set: according to initial to described image in described step 22
The packet that attitude data collection P is carried out, is divided into multiple described air strips number by described image data collection E
According to collection S1,S2,S3,...,SN, each described air strips data set corresponding initial appearance of described image respectively
One packet of state data set P, thus constitute described image setDescribed image set
The bridge image joint side of distinguished point based the most according to claim 1 and structure lines
Method, it is characterised in that described step 3 specifically includes following steps:
Step 31: divide image data collection: by each described air strips data set Sk, k=
1,2 ... N is further divided into multiple image data collection G1,G2,G3,...,GB, each described image
Data set Gk, k=1,2 ... B comprises the image of moderate number respectively, thus constitutes image set δ,
Described image set
Step 32: projection packet panorama sketch: for each described image in described image set δ
Data set Gk, k=1,2 ... B, carry out respectively registrating and projecting composition by its interior image data
Group panorama sketch Uk, k=1,2 ... B, thus constitute image set τ, described image set
Step 33: project individual air strips panorama sketch: by complete for the described packet in described image set τ
Scape figure UkRegistrate, project into individual described air strips panorama sketch Wk；
Step 34: project complete air strips panorama sketch: repeating said steps 31 arrives described step 33,
By described image setIn each described air strips data set SkRegistrate respectively, project into complete
Described air strips panorama sketch Wk, k=1,2 ... N, thus constitute described image set η, described image
Collection
The bridge image joint side of distinguished point based the most according to claim 3 and structure lines
Method, it is characterised in that described step 32 specifically includes following steps:
Step 321: feature point extraction and extraction of structure lines: to each described image data collection Gk,
K=1,2 ... B carries out feature point extraction and extraction of structure lines respectively, and wherein feature point extraction selects
SIFT feature point and ShiTomasi characteristic point；Extraction of structure lines then utilizes described bridge 3D point
The 3D structure lines that cloud data set M extracts, by the demarcation between camera, finds 3D to tie
2D structure lines on image corresponding to structure line；
Step 322: pyramid image mates: set up image pyramid, utilizes at described image
The SIFT feature point that pyramid is high-rise, the list that coupling is set up between image should be related to；By institute
State the list between image should be related to, on low layer pyramid, utilize ShiTomasi feature to click on
Row similarity measure mates, thus obtains precision preferable Image Matching point；
Described estimate matching formula and be expressed as:
Wherein,WithIt it is the m × n window chosen centered by characteristic pointWithThe meansigma methods of middle pixel, its computing formula is as follows:
Wherein, I (q) and I ' (q ') is the m × n window chosen centered by characteristic pointWithThe value of middle pixel, ρ (p, p ') is closer to 1, then the similarity between pixel is the highest,
Mate the most reliable；
The described Image Matching point obtained by Image Matching, thus set up described image data collection
GkCorresponding relation between interior image and image；
Step 323: obtain optimum image attitude: described in obtaining in described step 321
The described Image Matching point obtained in 2D structure lines and described step 322, more smart to obtain
True image attitude；
Energy-optimised function includes El(θ) and Ep(θ) two:
E (θ)=arg min (El(θ)+Ep(θ)),
Wherein El(θ) and Ep(θ) energy term of described 2D structure lines and described image are represented respectively
The energy term of match point,
Energy term E of described 2D structure linesl(θ) be calculated as follows,
Wherein, p is a series of value point such as grade on the described 2D structure lines of this image,Generation
Table is the attitude parameter of this image,Wait value spot projection to described 3D bridge for described
3D point on beam model,For distance on described 3D structure linesNearest 3D
Point；
Energy term E of described Image Matching pointp(θ) be calculated as follows,
Wherein, what I represented is coupling image, and what x and x ' represented is the match point on coupling image
Collection,WithRepresent is that the described Image Matching spot projection on coupling image is to institute
State spatial point corresponding on 3D bridge model；
By using LM function optimization, the initial attitude of image is adjusted, until function
Optimize energy term minimum, thus the optimum image attitude obtained；
Step 324: project individual packet panorama sketch: utilize the image attitude after optimizing by described
Image data collection GkProject on described 3D bridge model, obtain a described packet panorama sketch
Uk。
Step 325: projection is all grouped panorama sketch: repeating said steps 321 arrives described step
324, each packet in described image set τ is registrated respectively, and projects into complete described packet
Panorama sketch Uk, k=1,2 ... B, thus constitute described image set τ, described image set
The bridge image joint side of distinguished point based the most according to claim 3 and structure lines
Method, it is characterised in that described step 33 specifically includes following steps:
Step 331: setting up packet list should be related to: utilize adjacent described packet panorama sketch UkIt
Between the common region that has, set up described packet panorama sketch UkBetween list should be related to, and lead to
Cross described packet panorama sketch UkBetween list should be related to, by adjacent described packet panorama sketch Uk
Merge；
Step 332: project individual air strips panorama sketch: utilize the institute that described step 331 is set up
State packet panorama sketch UkBetween list should be related to, by each described packet panorama of described image set τ
Figure UkMerge one by one, the described air strips panorama sketch W of final acquisitionk。
The bridge image joint side of distinguished point based the most according to claim 1 and structure lines
Method, it is characterised in that described step 4 specifically includes following steps:
Step 41: set up air strips corresponding relation: by described air strips data set SkEach interior image,
Respectively with adjacent described air strips data set SkThe image inside with degree of overlapping carries out correspondence, thus
Set up adjacent air strips corresponding relation；
Step 42: set up air strips and singly should be related to: utilize described adjacent air strips corresponding relation, build
Vertical adjacent air strips singly should be related to；
Step 43: project adjacent air strips panorama sketch: utilize described adjacent air strips singly should be related to,
By panorama sketch W in air strips described inkProject to another adjacent described air strips panorama sketch WkOn；
Step 44: panorama sketch bottom projection bridge: repeating said steps 41 to described step 43,
Constantly by panorama sketch W in air strips described inkProject to another adjacent described air strips panorama sketch Wk
On, thus set up panorama sketch O bottom complete described bridge.
The bridge image joint side of distinguished point based the most according to claim 6 and structure lines
Method, it is characterised in that described step 41 specifically includes following steps:
Step 411: scaling: to described air strips data set SkInterior each image zooms in and out, and carries
Take its characteristic point, and carry out feature description, it is thus achieved that feature descriptor；
Step 412: characteristic matching: by described air strips data set SkInterior each image, respectively
With adjacent described air strips data set SkAll of image carries out characteristic matching one by one, i.e. calculates institute
Stating the Euclidean distance between feature descriptor, the point of selected distance minimum is as optimal match point；
Step 413: set up air strips corresponding relation: the number according to described optimal match point is how many
Select optimal image to match, thus set up described adjacent air strips corresponding relation.
The bridge image joint side of distinguished point based the most according to claim 6 and structure lines
Method, it is characterised in that described step 42 specifically includes following steps:
Step 421: air strips feature point extraction: utilize described adjacent air strips data set corresponding relation,
Carrying out feature point extraction wherein on a pair image respectively, feature point extraction uses SIFT
Characteristic point and ShiTomasi characteristic point；
Step 422: air strips pyramid image mates: on the image of described step 421, weight
The operation of multiple described step 322, thus obtain precision preferable air strips data set Image Matching point；
Step 423: set up air strips and singly should be related to: utilize described air strips data set Image Matching point,
Set up described adjacent air strips by matching double points singly should be related to.
The bridge image joint side of distinguished point based the most according to claim 1 and structure lines
Method, it is characterised in that described step 5 specifically includes following steps:
Step 51: image set is grouped: be grouped all image set being registered,
Moderate image number is comprised in making often to organize；
Step 52: image joint: to the image in each packet, carries out splicing line simultaneously and looks into
Look for and processing with light and color homogenization, form a big figure；
Step 53: packet splicing: the big figure to the formation of each packet, splices again
Line is searched and light and color homogenization processes, and ultimately forms image panorama sketch at the bottom of a complete bridge.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610269079.7A CN105957005B (en) | 2016-04-27 | 2016-04-27 | Bridge image splicing method based on characteristic point and structure lines |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610269079.7A CN105957005B (en) | 2016-04-27 | 2016-04-27 | Bridge image splicing method based on characteristic point and structure lines |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105957005A true CN105957005A (en) | 2016-09-21 |
| CN105957005B CN105957005B (en) | 2019-03-29 |
Family
ID=56915749
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610269079.7A Active CN105957005B (en) | 2016-04-27 | 2016-04-27 | Bridge image splicing method based on characteristic point and structure lines |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105957005B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107316299A (en) * | 2017-07-13 | 2017-11-03 | 云南数云信息科技有限公司 | Ancient architecture deformed monitoring method and system based on three-dimensional point cloud technology |
| CN107369211A (en) * | 2017-07-13 | 2017-11-21 | 云南数云信息科技有限公司 | Ancient building three-dimensional point cloud acquisition system and model review one's lessons by oneself correction method |
| CN107392964A (en) * | 2017-07-07 | 2017-11-24 | 武汉大学 | The indoor SLAM methods combined based on indoor characteristic point and structure lines |
| CN112669388A (en) * | 2019-09-30 | 2021-04-16 | 上海禾赛科技股份有限公司 | Calibration method and device for laser radar and camera device and readable storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080247635A1 (en) * | 2006-03-20 | 2008-10-09 | Siemens Power Generation, Inc. | Method of Coalescing Information About Inspected Objects |
| CN101592484A (en) * | 2009-07-09 | 2009-12-02 | 浙江省电力公司 | A kind of distance-finding method of line of electric force |
| US20090300692A1 (en) * | 2008-06-02 | 2009-12-03 | Mavlankar Aditya A | Systems and methods for video streaming and display |
| CN101916452A (en) * | 2010-07-26 | 2010-12-15 | 中国科学院遥感应用研究所 | Method for automatically stitching unmanned aerial vehicle remote sensing images based on flight control information |
| US20120076358A1 (en) * | 2004-08-31 | 2012-03-29 | Meadow William D | Methods for and Apparatus for Generating a Continuum of Three-Dimensional Image Data |
| CN104363438A (en) * | 2014-12-03 | 2015-02-18 | 中测新图(北京)遥感技术有限责任公司 | Panoramic three-dimensional image manufacturing method |
-
2016
- 2016-04-27 CN CN201610269079.7A patent/CN105957005B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120076358A1 (en) * | 2004-08-31 | 2012-03-29 | Meadow William D | Methods for and Apparatus for Generating a Continuum of Three-Dimensional Image Data |
| US20080247635A1 (en) * | 2006-03-20 | 2008-10-09 | Siemens Power Generation, Inc. | Method of Coalescing Information About Inspected Objects |
| US20090300692A1 (en) * | 2008-06-02 | 2009-12-03 | Mavlankar Aditya A | Systems and methods for video streaming and display |
| CN101592484A (en) * | 2009-07-09 | 2009-12-02 | 浙江省电力公司 | A kind of distance-finding method of line of electric force |
| CN101916452A (en) * | 2010-07-26 | 2010-12-15 | 中国科学院遥感应用研究所 | Method for automatically stitching unmanned aerial vehicle remote sensing images based on flight control information |
| CN104363438A (en) * | 2014-12-03 | 2015-02-18 | 中测新图(北京)遥感技术有限责任公司 | Panoramic three-dimensional image manufacturing method |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107392964A (en) * | 2017-07-07 | 2017-11-24 | 武汉大学 | The indoor SLAM methods combined based on indoor characteristic point and structure lines |
| CN107392964B (en) * | 2017-07-07 | 2019-09-17 | 武汉大学 | The indoor SLAM method combined based on indoor characteristic point and structure lines |
| CN107316299A (en) * | 2017-07-13 | 2017-11-03 | 云南数云信息科技有限公司 | Ancient architecture deformed monitoring method and system based on three-dimensional point cloud technology |
| CN107369211A (en) * | 2017-07-13 | 2017-11-21 | 云南数云信息科技有限公司 | Ancient building three-dimensional point cloud acquisition system and model review one's lessons by oneself correction method |
| CN112669388A (en) * | 2019-09-30 | 2021-04-16 | 上海禾赛科技股份有限公司 | Calibration method and device for laser radar and camera device and readable storage medium |
| CN112669388B (en) * | 2019-09-30 | 2022-06-21 | 上海禾赛科技有限公司 | Calibration method and device for laser radar and camera device and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105957005B (en) | 2019-03-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112894832B (en) | Three-dimensional modeling method, three-dimensional modeling device, electronic equipment and storage medium | |
| CN111968129B (en) | Instant positioning and map construction system and method with semantic perception | |
| CN111209915B (en) | Three-dimensional image synchronous recognition and segmentation method based on deep learning | |
| CN110110641B (en) | Unmanned aerial vehicle monitoring method and system for watershed flooding scene | |
| CN109165680B (en) | Single-target object dictionary model improvement method in indoor scene based on visual SLAM | |
| CN103927731B (en) | Low-altitude remote sensing image rapid and automatic splicing method without POS assisting | |
| CN109509230A (en) | A kind of SLAM method applied to more camera lens combined type panorama cameras | |
| CN113362247B (en) | Semantic real scene three-dimensional reconstruction method and system for laser fusion multi-view camera | |
| CN113409459B (en) | Method, device and equipment for producing high-precision map and computer storage medium | |
| CN105957005A (en) | Method for bridge image splicing based on feature points and structure lines | |
| CN110244766A (en) | A kind of method and system for planning in photovoltaic plant unmanned plane inspection course line | |
| CN112598729B (en) | Target object identification and positioning method integrating laser and camera | |
| CN101394573A (en) | Panoramagram generation method and system based on characteristic matching | |
| CN105096283A (en) | Panoramic image acquisition method and device | |
| CN110660125B (en) | Three-dimensional modeling device for power distribution network system | |
| Ma et al. | Crlf: Automatic calibration and refinement based on line feature for lidar and camera in road scenes | |
| CN112288637A (en) | Unmanned aerial vehicle aerial image rapid splicing device and rapid splicing method | |
| CN117036300A (en) | Road surface crack identification method based on point cloud-RGB heterogeneous image multistage registration mapping | |
| CN114842139A (en) | Building three-dimensional digital model construction method based on spatial analysis | |
| CN107958489B (en) | Curved surface reconstruction method and device | |
| CN109035365B (en) | Mosaic processing method of high-resolution image | |
| CN114581307A (en) | Multi-image stitching method, system, device and medium for target tracking identification | |
| CN110160503A (en) | A kind of unmanned plane landscape matching locating method for taking elevation into account | |
| CN111986320B (en) | Smart city application-oriented DEM and oblique photography model space fitting optimization method | |
| CN116740288A (en) | Three-dimensional reconstruction method integrating laser radar and oblique photography |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |