CN110033411A - The efficient joining method of highway construction scene panoramic picture based on unmanned plane - Google Patents
The efficient joining method of highway construction scene panoramic picture based on unmanned plane Download PDFInfo
- Publication number
- CN110033411A CN110033411A CN201910292872.2A CN201910292872A CN110033411A CN 110033411 A CN110033411 A CN 110033411A CN 201910292872 A CN201910292872 A CN 201910292872A CN 110033411 A CN110033411 A CN 110033411A
- Authority
- CN
- China
- Prior art keywords
- image
- point
- unmanned plane
- pixel
- circle
- 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
- 238000010276 construction Methods 0.000 title claims abstract description 34
- 238000000034 method Methods 0.000 title claims abstract description 27
- 238000006243 chemical reaction Methods 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 claims description 7
- 239000011159 matrix material Substances 0.000 claims description 7
- 238000002156 mixing Methods 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 6
- 230000002093 peripheral effect Effects 0.000 claims description 6
- 230000004927 fusion Effects 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 5
- 238000013519 translation Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 3
- 238000009416 shuttering Methods 0.000 claims description 3
- 230000008901 benefit Effects 0.000 abstract description 3
- 230000006870 function Effects 0.000 description 7
- 238000007796 conventional method Methods 0.000 description 6
- 238000007726 management method Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 5
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 238000001514 detection method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000012876 topography Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 230000000007 visual effect Effects 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
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The present invention relates to a kind of, and the efficient joining method of highway construction scene panoramic picture based on unmanned plane is corrected by the geography information coordinate and attitude parameter of Aerial Images, crucial splicing regions are chosen, characteristic point efficient matchings, fast image splicing based on optimal stitching line and image co-registration, solves the local coordinate system deviation that unmanned plane generates during cruise, entire image Feature Points Matching is inefficient, and since the splicing that dynamic object generates is fuzzy and ghost problem, with real time transmitting image data, overall situation supervision, accuracy is good, it is high-efficient, it is at low cost, the advantages that convenient, flexible, to realize entire construction site automation, intelligentized global safety supervision and management is laid a good foundation.The present invention is suitable for the global safety supervision and management at highway engineering construction scene.
Description
Technical field
The present invention relates to a kind of efficient joining method of highway construction scene panoramic picture based on unmanned plane.
Background technique
With the quickening of China's transport development development process, the Construction quality and safety thing of traffic engineering especially highway engineering
Therefore also present easily hair, multiple, high-incidence situation.Currently, the safety supervision management at highway engineering construction scene mostly uses people
Work telescope, the installation traditional means such as camera, that there is independences is poor, flexibility ratio is low, observation area is restricted and there is sight
The disadvantages of surveying blind area, global planning's management big by the influence of topography, can not achieve construction site.For above-mentioned problem, both at home and abroad
Scholar has carried out the engineering safety management method research of view-based access control model.As the image acquisition hardwares such as unmanned plane are set in recent years
The development of the software algorithms technologies such as standby and computer vision, image procossing has had some based on unmanned plane and image spelling at present
The generation method of the construction site panoramic picture connect, to realize that early period has been established in the global safety supervision and management of entire construction site
Research foundation.However, these methods can not often obtain really effective application in practical projects.To find out its cause, first
It is the deviation that unmanned plane can generate local coordinate system during cruise, such as the visual angle fluctuation in plane of cruising, conventional method does not have
There are the geography information coordinate for considering unmanned plane and attitude parameter amendment, so as to cause topography's distortion, and then causes generating
Stitching error when panorama sketch is fairly obvious;Secondly because current merging algorithm for images is all based on entire image area mostly
What the Feature Points Matching in domain carried out, treatment effeciency is lower, reality of the Yao Shixian in several live high-definition pictures of practice of construction
When or quasi real time quickly splicing be very difficult;Furthermore due to inevitably by natural wind effect and construction site
There is fuzzy and ghost in the spliced panoramic figure that conventional method generates more in the track demand of dynamic object.How nobody is directed to
Position and attitudes vibration of the machine during cruise propose a kind of efficient, accurate Panorama Mosaic method, be one urgently
It solves the problems, such as.
Summary of the invention
Based on the above shortcomings, the present invention proposes that a kind of highway construction scene panoramic picture based on unmanned plane is efficiently spelled
Connect method, solve the local coordinate system deviation that unmanned plane generates during cruise, entire image Feature Points Matching it is inefficient,
And since the splicing that dynamic object generates is fuzzy and ghost problem.
The technology used in the present invention is as follows: a kind of efficient side of splicing of the highway construction scene panoramic picture based on unmanned plane
Method, steps are as follows:
Extracted Step 1: carrying out geography information and attitude parameter to unmanned plane acquired image, based on gauss projection and
Coordinate rotation translation realizes unmanned plane location information by the conversion of geographic coordinate system to local coordinate system, and according to unmanned plane appearance
State parameter carries out homography matrix amendment, eliminates image fault error caused by wind-induced vibration cruise angular variation;
Step 2: will be carried out by geography information and the revised image of attitude parameter it is adjacent match two-by-two, based on part
Pixel changes the crucial splicing regions in part of maximum value selected characteristic point, and carries out the feature in key area based on ORB feature
Point matching;
Step 3: based on color and geometric error minimum principle and overlapping region weighted average, iteration carries out adjacent image
Optimal stitching line search and image boundary sectionally weighting blending algorithm, eliminate that splicing is fuzzy and ghost phenomenon, obtain final
Panoramic mosaic image.
The present invention also has following technical characteristic:
1, step 1 as described above specifically includes:
Step 1 one, using winged control platform courses unmanned plane during flying direction and speed, guarantee that the Duplication of adjacent image is protected
Card realizes the continuous processing of multiple image 50%;
Step 1 two, the image obtained to step 1 one carry out serial number, and by the geography information extracted and posture
Parameter is modified, and carries out the conversion of multiple image homography matrix and registration;
Wherein, the positive calculation formula of Gauss-Ke Lvke projection of the plane rectangular coordinates and geographical coordinate of coordinate conversion is as follows:
Wherein, x, y are the transverse and longitudinal coordinate of plane right-angle coordinate, and L, B are the longitude and latitude of ellipsoid geographic coordinate system, and s is
Meridional parts from equator, N, η are respectively radius of curvature and intermediate variable, and calculation formula is as follows:
Wherein, a is that semimajor axis of ellipsoid is long, e, e ' it is respectively the first, second ellipticity of ellipsoid.
2, in step 2 as described above, the choosing method of crucial splicing regions are as follows:
Wherein, x, y have respectively represented coordinate value of the pixel on width and short transverse, and I indicates gray value of image, Ii∩
Ii+1Indicate the overlapping region of the i-th width and i+1 width image, Duplication is by flying control platform courses 50%.
3, in step 2 as described above, after completing crucial splicing regions and choosing, wherein the extraction process of ORB feature is such as
Under: first using any pixel on image as the center of circle, circle is done on the image with radii fixus, counts the pixel that peripheral circular arc passes through
Gray value, then the gray value of more peripheral circular arc pixel and central point pixel, statistics gray scale difference value are greater than of given threshold
Number, and in this, as judge central pixel point whether be candidate feature point foundation, circular shuttering radius be 3 pixels, will be to be checked
Pixel compares in the circle that measuring point p is constituted with surrounding 16 pixels, judges in the circle with the presence or absence of enough
Pixel is different from p attribute, if it is present p is angle point, in gray level image, algorithm be by the gray value of each point and p point into
Row compare, if there is n continuous image vegetarian refreshments ratio p point out it is all bright or all secretly, then p is angle point, n=9, then in key point p
Around N number of point pair chosen with certain pattern, the comparison result of this N number of point pair in combination as description.With key point p
For the center of circle, it is round O by radius of d, a certain mode chooses N number of point pair in circle O, and N can take 512, using key point as the center of circle, with
Key point and the line for taking the mass center in region are that X-axis establishes two-dimensional coordinate system, when two o'clock similarity is greater than threshold value, then two o'clock
Successful match.
4, in step 3 as described above, the objective optimization function of optimal stitching line specifically:
E (x, y)=Ecolor(x,y)2+Egeometry(x,y) (4)
Ecolor=Δ Ii=Ii+1-Ii (6)
In formula, E indicates the objective optimization function of optimal stitching line, EcolorIndicate superposition image vegetarian refreshments on two width original images
The difference of color value, EgeometryIndicate the structure difference of superposition image vegetarian refreshments on two width original images.Sx,SyRespectively indicate Sobel ladder
Spend operator, Ii,Ii+1Two width adjacent images are respectively represented,Indicate convolution algorithm.
5, image boundary sectionally weighting blending algorithm in step 3 as described above, the specific steps are as follows:
In formula, (x, y) ∈ R indicates that the pixel in crucial splicing regions R, f (x, y) represent the image after Weighted Fusion,
fi(x, y) represents the i-th width original image, and i=1,2 indicate continuous two adjacent images splice, di(x, y) is indicated
The sectionally weighting weight coefficient changed as location of pixels changes, changes linearly, value range 0 along picture altitude direction
~1, h are the picture altitude of crucial splicing regions.Wherein, computation rate is also resided in using the advantage of sectionally weighting blending algorithm
Fastly, and explicit physical meaning.In each image to be spliced, apart from upper width image remote position (y is closer to h),
With it is closer at a distance from optimal stitching line, corresponding weight coefficient is played the role of also bigger closer to 1 in image co-registration.
After completing image co-registration near optimal stitching line, the splicing of conventional method bring can be eliminated and obscured and ghost phenomenon.
The invention has the benefit that being directed to local coordinate system deviation, the whole picture figure that unmanned plane generates during cruise
As Feature Points Matching is inefficient and since the splicing that dynamic object generates obscures and ghost problem, passes through the ground of Aerial Images
It manages information coordinate and attitude parameter amendment, crucial splicing regions selection and characteristic point efficient matchings, be based on optimal stitching line and figure
As the fast image splicing of fusion, the splicing of construction site unmanned plane panorama high-definition picture is realized.This method improves nothing
The accuracy of the computational efficiency and splicing result of man-machine panorama high-definition picture splicing, significantly reduces in conventional method
Artificial participation.The present invention is also able to satisfy the on-line monitoring early warning and real time data processing demand of construction site, directly right
Acquired image carries out transmission splicing, and as a result output delay can be down to ten seconds or less.The present invention improves the construction site overall situation
Automation, intelligence degree and the accuracy of safety supervision management are the global safety supervision and management at Traffic Engineering Construction scene
Provide solution.
Detailed description of the invention
Fig. 1 is the flow chart of one embodiment of the present of invention;
Fig. 2 is the flow diagram of core algorithm of the present invention;
Fig. 3 is that the key area of step 2 in the present invention chooses result figure;
Fig. 4 is that result figure is matched in ORB characteristic point area in the key area of step 2 in the present invention;
Fig. 5 is the optimal stitching line result figure of step 3 in the present invention, and wherein black broken line represents adjacent two images
Optimal stitching line;
Fig. 6 is the global high definition splicing result figure in highway engineering construction scene that the present invention carries out embodiment;
Fig. 7 is that elimination of the invention is fuzzy with ghost effect picture, the partial splice that wherein Fig. 7 (a) generates for conventional method
Fuzzy and ghost figure, Fig. 7 (b) are the high-resolution result figure that the present invention generates.
Specific embodiment
Embodiment 1
Present embodiment is based on unmanned plane geography information and the modified highway engineering construction scene panorama sketch of attitude parameter
Image height imitates joining method, as shown in Figure 1, comprising:
Extracted Step 1: carrying out geography information and attitude parameter to unmanned plane acquired image, based on gauss projection and
Coordinate rotation translation realizes unmanned plane location information by the conversion of geographic coordinate system to local coordinate system, and according to unmanned plane appearance
State parameter carries out homography matrix amendment, eliminates image fault error caused by wind-induced vibration cruise angular variation.
For example, in one embodiment, the resolution ratio of single width original color image is 5472 X 3684, original image is mentioned
Take out the attitude parameters such as geography information and pitch angle, course angle, roll angle such as longitude, latitude, elevation of corresponding camera site.
Then it is matched two-by-two according to adjacent image, carries out the conversion of multiple image homography matrix, obtain the continuous registration knot of multiple image
Fruit.
Step 2: will be carried out by geography information and the revised image of attitude parameter it is adjacent match two-by-two, based on part
Pixel changes the crucial splicing regions in part of maximum value selected characteristic point, and is based on ORB (Oriented FAST and
Rotated BRIEF spins up segmentation and binaryzation robust separate unit) feature carry out key area in characteristic point
Match.
Wherein, the selection of the crucial splicing regions in part is the overlapping region based on image to be spliced.Crucial splicing regions
Height be 50% picture altitude, i.e. 1842 pixels;It is 4- that width, which is then according to the tilt angle of analysis image path breadths edge,
6 °, it is 125~190 that pixel number is laterally accounted in half image, so automatic frame selects each 150 pixel numbers in left and right when selecting;Due to
The influence conditions such as construction site wind speed make image lateral run-out, and deviation value is within 100 pixels, so when neighbor map frame selects
Target area is chosen in 100 pixels of each increase in left and right, guarantee.Crucial splicing regions width is 300 pixels, back panel in i.e. preceding width image
It is 500 pixels in image, guarantees the Feature Points Matching effect in crucial splicing regions.Fig. 3 is that the key area of adjacent image selects
Result figure is taken, Fig. 4 is that result figure is matched in ORB characteristic point area in key area.
Step 3: based on color and geometric error minimum principle and overlapping region weighted average, iteration carries out adjacent image
Optimal stitching line search and image boundary fusion, eliminate that splicing is fuzzy and ghost phenomenon, obtain final panoramic mosaic image.
The process of search optimal stitching line is according to color and geometric error minimum principle, by the part of two images overlapping
Make difference operation and generates a width error image;Then the first row of the thought of Dynamic Programming from overlapping region is used to this error image
It sets out, establishes using each pixel on the row as the suture of starting point;An optimal seam is finally found from these sutures
Zygonema.Specific steps are as follows: initialization each column pixel of the first row corresponds to a suture, and intensity value is initialized as each point
Criterion value, the current point of the suture is the train value where it;A line that extension had been computed suture intensity expands downwards
Exhibition, to the last until a line, the method for extension is by 3 in the current point of each suture and the adjacent next line of point
The addition of a pixel criterion value is compared, take this 3 pixels of next line corresponding to minimal intensity value one of be used as the seam
The propagation direction of zygonema, the intensity value for updating this suture is minimal intensity value, and the current point of suture is updated to obtain
The column where adjacent pixels value in next line where minimal intensity value;Optimal stitching line is selected, from all sutures
Selection intensity value is the smallest to be used as optimal stitching line.The dimension of picture for inputting the picture for inputing to model when meeting trained.Fig. 5
In black broken line just represent the optimal stitching line of adjacent two images.
The operation result of embodiment is developed under 2.0 environment of MATLAB 2016a and OpenCV, is directly applied for using
The construction site image of consumer level unmanned plane shooting, does not need special shooting or detection device, and cruising altitude is 30 meters.We
Method splices precision height, and speed is fast, at low cost, can be not only used for the identified off-line of construction site global safety assessment, it can also be used to quasi-
Real-time monitoring handled time delay within 5 seconds, improved the automation of construction site global safety supervision, intelligence, accurate
Property and treatment effeciency.
Fig. 6 to Fig. 7 is the splicing effect figure of the embodiment of the present invention, and wherein Fig. 6 is that continuously spliced high definition is complete for 8 width images
Jing Tu, Fig. 7 (a) are the fuzzy high-resolution generated with ghost figure, Fig. 7 (b) for the present invention of partial splice that conventional method generates
Result figure.
Embodiment 2
Present embodiment is substantially the same with 1 step of embodiment, the difference is that: step 1 specifically includes:
Step 1 one, using winged control platform PIX4D, control unmanned plane during flying direction and speed, guarantee the weight of adjacent image
Folded rate guarantees to realize the continuous processing of multiple image 50%.
Step 1 two, the image obtained to step 1 one carry out serial number, and by the geography information extracted and posture
Parameter is modified, and carries out the conversion of multiple image homography matrix and registration.
Wherein, the positive calculation formula of Gauss-Ke Lvke projection of the plane rectangular coordinates and geographical coordinate of coordinate conversion is as follows:
Wherein, x, y are the transverse and longitudinal coordinate of plane right-angle coordinate, and L, B are the longitude and latitude of ellipsoid geographic coordinate system, and s is
Meridional parts from equator, N, η are respectively radius of curvature and intermediate variable, and calculation formula is as follows:
Wherein, a is that semimajor axis of ellipsoid is long, e, e ' it is respectively the first, second ellipticity of ellipsoid,
The beneficial effect of present embodiment is to realize local coordinate conversion by extracting unmanned plane geographic coordinate information;
And ensure that the Duplication of adjacent image is 50% by flying control platform, the continuous spelling of construction site global image may be implemented
It connects.
It has collected the image that 8 width Duplication are 50% to a high road pavement construction section construction site cover entirely
Lid is projected using Gauss-Ke Lvke, and geographical coordinate is projected to plane coordinate system and carries out relative position calculating.In embodiment,
The 8 width consecutive images that, Duplication collected to unmanned plane is 50% carry out geography information and attitude parameter extracts, and based on height
This projection and coordinate rotation translation, realize unmanned plane geography information and attitude parameter amendment, the results are shown in Table 1.
1 geography information of table and attitude parameter correction result
Other steps are same as Example 1.
Embodiment 3
Present embodiment is substantially the same with 1 step of embodiment, the difference is that: in step 2, the choosing of key area
The principle is taken to be
Wherein, x, y have respectively represented coordinate value of the pixel on width and short transverse, and I indicates gray value of image, Ii∩
Ii+1Indicate the overlapping region of the i-th width and i+1 width image, Duplication is by flying control platform courses 50%.
After completing key area and choosing, wherein the extraction process of ORB feature is as follows:
Based on FAST Corner Detection and BRIEF Feature Descriptor, ORB feature has good robustness and real-time, and
And calculating cost and memory needs are very low.First using any pixel on image as the center of circle, circle is done on the image with radii fixus,
The gray value for the pixel that peripheral circular arc passes through is counted, then the gray value of more peripheral circular arc pixel and central point pixel, statistics
Gray scale difference value be greater than given threshold number, and in this, as judge central pixel point whether be candidate feature point foundation.Often
Circular shuttering radius is 3 pixels, and pixel in circle that measuring point p to be checked is constituted with surrounding 16 pixels is done ratio
Compared with, judge it is different from p attribute with the presence or absence of enough pixels in the circle, if it is present p may be angle point, in gray scale
In image, algorithm is to be compared the gray value of each point with p point, if there is n continuous image vegetarian refreshments ratio p point out it is all bright or
All dark, then p may be angle point.By test, n=9, treatment effect, rate and the robustness that algorithm obtains are all very good.Then
N number of point pair is chosen with certain pattern around key point P, the comparison result this N number of point pair is sub in combination as description.
Using key point P as the center of circle, it is round O by radius of d, a certain mode chooses N number of point pair in circle O.N can take in practical application
512.Using key point as the center of circle, two-dimensional coordinate system is established as X-axis using the line of key point and the mass center for taking a region.Different
It rotates under angle, is consistent with the same point for taking dot pattern to take out, to solve the problems, such as rotation consistency.Last root
According to Feature Descriptor given threshold, such as A:10101011, B:10101010, when two o'clock similarity is greater than threshold value, then two o'clock
Successful match.
Other steps and parameter are same as Example 1.
Embodiment 4
Present embodiment is substantially the same with 1 step of embodiment, the difference is that: in step 3, optimal stitching line
Objective optimization function specifically:
E (x, y)=Ecolor(x,y)2+Egeometry(x,y) (4)
Ecolor=Δ Ii=Ii+1-Ii (6)
In formula, E indicates the objective optimization function of optimal stitching line, EcolorIndicate superposition image vegetarian refreshments on two width original images
The difference of color value, EgeometryIndicate the structure difference of superposition image vegetarian refreshments on two width original images.Sx,SyRespectively indicate Sobel ladder
Spend operator, Ii,Ii+1Two width adjacent images are respectively represented,Indicate convolution algorithm.
Other steps and parameter are identical as tool embodiment 1.
Embodiment 5
Present embodiment is substantially the same with 1 step of embodiment, the difference is that: segmentation, which is used, in step 3 adds
Weigh blending algorithm.The reason is that if stitching portion can generate apparent splicing seams in splicing by image simple superposition, in order to
It eliminates splicing seams and introduces the sectionally weighting blending algorithm of image.The selection for being weighted and averaged weight function gradually goes out method using progressive,
Using pixel to the Euclidean distance at overlapping region center as weight function.When image is in overlay region transition, weight function is by 1
It is gradient to 0.This Weighted Average Algorithm can handle difference in exposure well, and fusion speed is fast, realize that simply real-time is good.
In hardware configuration be 8GB DDR3 memory and i7-4790CPU, software environment are the consumption of MATLAB 2016a and OpenCV 2.0
The processing time on grade laptop is about 4s, and tradition is about based on the greedy SIFT feature matched image split-joint method used time
13s, efficiency improve nearly 2 times.
Other steps and parameter are identical as tool embodiment 1.
Claims (6)
1. a kind of efficient joining method of highway construction scene panoramic picture based on unmanned plane, which is characterized in that steps are as follows:
Step 1: carrying out geography information and attitude parameter extraction to unmanned plane acquired image, it is based on gauss projection and coordinate
Rotation translation realizes that unmanned plane location information by the conversion of geographic coordinate system to local coordinate system, and is joined according to UAV Attitude
Number carries out homography matrix amendment, eliminates image fault error caused by wind-induced vibration cruise angular variation;
Step 2: will be carried out by geography information and the revised image of attitude parameter it is adjacent match two-by-two, be based on local pixel
Change the crucial splicing regions in part of maximum value selected characteristic point, and the characteristic point in key area is carried out based on ORB feature
Match;
Step 3: based on color and geometric error minimum principle and overlapping region weighted average, iteration carries out adjacent image most
Good suture line search and image boundary sectionally weighting blending algorithm are eliminated and splice fuzzy and ghost phenomenon, obtain final panorama
Stitching image.
2. the highway construction scene panoramic picture efficient joining method according to claim 1 based on unmanned plane, feature
It is, step 1 specifically includes:
Step 1 one, using winged control platform courses unmanned plane during flying direction and speed, guarantee that the Duplication of adjacent image guarantees
50%, realize the continuous processing of multiple image;
Step 1 two, the image obtained to step 1 one carry out serial number, and by the geography information extracted and attitude parameter
It is modified, and carries out the conversion of multiple image homography matrix and registration;
Wherein, the positive calculation formula of Gauss-Ke Lvke projection of the plane rectangular coordinates and geographical coordinate of coordinate conversion is as follows:
Wherein, x, y are the transverse and longitudinal coordinate of plane right-angle coordinate, and L, B are the longitude and latitude of ellipsoid geographic coordinate system, and s is from red
The meridional parts that road rises, N, η are respectively radius of curvature and intermediate variable, and calculation formula is as follows:
Wherein, a is that semimajor axis of ellipsoid is long, e, e ' it is respectively the first, second ellipticity of ellipsoid.
3. the highway construction scene panoramic picture efficient joining method according to claim 1 or 2 based on unmanned plane, special
Sign is, in step 2, the choosing method of crucial splicing regions are as follows:
Wherein, x, y have respectively represented coordinate value of the pixel on width and short transverse, and I indicates gray value of image, Ii∩Ii+1Table
Show the overlapping region of the i-th width and i+1 width image, Duplication is by flying control platform courses 50%.
4. the highway construction scene panoramic picture efficient joining method according to claim 1 or 2 based on unmanned plane, special
Sign is, in step 2, after completing crucial splicing regions and choosing, wherein the matching process of ORB feature is as follows: first with image
Upper any pixel is the center of circle, does circle on the image with radii fixus, counts the gray value for the pixel that peripheral circular arc passes through, then compares
The gray value of more peripheral circular arc pixel and central point pixel, statistics gray scale difference value be greater than given threshold number, and in this, as
Judge central pixel point whether be candidate feature point foundation, circular shuttering radius be 3 pixels, by measuring point p to be checked with around it
The circle that is constituted of 16 pixels in pixel compare, judge in the circle with the presence or absence of enough pixel and p attribute
Difference, if it is present p is angle point, in gray level image, algorithm is to be compared the gray value of each point with p point, if deposited
Pointed out in n continuous image vegetarian refreshments ratio p all bright or all dark, then p is angle point, n=9, then with a cover half around key point p
Formula chooses N number of point pair, and the comparison result this N number of point pair is sub in combination as description.It is half with d using key point p as the center of circle
Diameter is round O, and a certain mode chooses N number of point pair in circle O, and N can take 512, using key point as the center of circle, with key point and the area Qu Dian
The line of the mass center in domain is that X-axis establishes two-dimensional coordinate system, when two o'clock similarity is greater than threshold value, then two o'clock successful match.
5. the highway construction scene panoramic picture efficient joining method according to claim 1 or 2 based on unmanned plane, special
Sign is, in step 3, the objective optimization function of optimal stitching line specifically:
E (x, y)=Ecolor(x,y)2+Egeometry(x,y) (4)
Ecolor=Δ Ii=Ii+1-Ii (6)
In formula, E indicates the objective optimization function of optimal stitching line, EcolorIndicate the color of superposition image vegetarian refreshments on two width original images
The difference of value, EgeometryIndicate the structure difference of superposition image vegetarian refreshments on two width original images, Sx,SyRespectively indicate the calculation of Sobel gradient
Son, Ii,Ii+1Two width adjacent images are respectively represented,Indicate convolution algorithm.
6. the highway construction scene panoramic picture efficient joining method according to claim 1 or 2 based on unmanned plane, special
Sign is, the image boundary sectionally weighting blending algorithm in step 3, the specific steps are as follows:
In formula, (x, y) ∈ R indicates that the pixel in crucial splicing regions R, f (x, y) represent the image after Weighted Fusion, fi(x,
Y) the i-th width original image is represented, i=1,2 indicate continuous two adjacent images splice, di(x, y) indicate with
The sectionally weighting weight coefficient of location of pixels variation and change, changes linearly along picture altitude direction, and value range is 0~1,
H is the picture altitude of crucial splicing regions.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910292872.2A CN110033411B (en) | 2019-04-12 | 2019-04-12 | High-efficiency road construction site panoramic image splicing method based on unmanned aerial vehicle |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910292872.2A CN110033411B (en) | 2019-04-12 | 2019-04-12 | High-efficiency road construction site panoramic image splicing method based on unmanned aerial vehicle |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110033411A true CN110033411A (en) | 2019-07-19 |
CN110033411B CN110033411B (en) | 2021-01-12 |
Family
ID=67238177
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910292872.2A Active CN110033411B (en) | 2019-04-12 | 2019-04-12 | High-efficiency road construction site panoramic image splicing method based on unmanned aerial vehicle |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110033411B (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110569927A (en) * | 2019-09-19 | 2019-12-13 | 浙江大搜车软件技术有限公司 | Method, terminal and computer equipment for scanning and extracting panoramic image of mobile terminal |
CN110796734A (en) * | 2019-10-31 | 2020-02-14 | 中国民航科学技术研究院 | Airport clearance patrol method and device based on high-resolution satellite technology |
CN111680703A (en) * | 2020-06-01 | 2020-09-18 | 中国电建集团昆明勘测设计研究院有限公司 | 360-degree construction panorama linkage positioning method based on image feature point detection and matching |
CN112184662A (en) * | 2020-09-27 | 2021-01-05 | 成都数之联科技有限公司 | Camera external parameter initial method and system applied to unmanned aerial vehicle image stitching |
CN112215304A (en) * | 2020-11-05 | 2021-01-12 | 珠海大横琴科技发展有限公司 | Gray level image matching method and device for geographic image splicing |
CN112308774A (en) * | 2020-09-15 | 2021-02-02 | 北京中科遥数信息技术有限公司 | Unmanned aerial vehicle-based map reconstruction method and system, transmission equipment and storage medium |
CN112907452A (en) * | 2021-04-09 | 2021-06-04 | 长春理工大学 | Optimal suture line searching method for image stitching |
CN113228105A (en) * | 2019-12-30 | 2021-08-06 | 商汤国际私人有限公司 | Image processing method and device and electronic equipment |
CN113286081A (en) * | 2021-05-18 | 2021-08-20 | 中国民用航空总局第二研究所 | Target identification method, device, equipment and medium for airport panoramic video |
CN117687426A (en) * | 2024-01-31 | 2024-03-12 | 成都航空职业技术学院 | Unmanned aerial vehicle flight control method and system in low-altitude environment |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
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 |
CN103426153A (en) * | 2013-07-24 | 2013-12-04 | 广州地理研究所 | Unmanned aerial vehicle remote sensing image quick splicing method |
CN106023086A (en) * | 2016-07-06 | 2016-10-12 | 中国电子科技集团公司第二十八研究所 | Aerial photography image and geographical data splicing method based on ORB feature matching |
CN107808362A (en) * | 2017-11-15 | 2018-03-16 | 北京工业大学 | A kind of image split-joint method combined based on unmanned plane POS information with image SURF features |
CN109389555A (en) * | 2018-09-14 | 2019-02-26 | 复旦大学 | A kind of Panorama Mosaic method and device |
-
2019
- 2019-04-12 CN CN201910292872.2A patent/CN110033411B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
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 |
CN103426153A (en) * | 2013-07-24 | 2013-12-04 | 广州地理研究所 | Unmanned aerial vehicle remote sensing image quick splicing method |
CN106023086A (en) * | 2016-07-06 | 2016-10-12 | 中国电子科技集团公司第二十八研究所 | Aerial photography image and geographical data splicing method based on ORB feature matching |
CN107808362A (en) * | 2017-11-15 | 2018-03-16 | 北京工业大学 | A kind of image split-joint method combined based on unmanned plane POS information with image SURF features |
CN109389555A (en) * | 2018-09-14 | 2019-02-26 | 复旦大学 | A kind of Panorama Mosaic method and device |
Non-Patent Citations (4)
Title |
---|
ETHAN RUBLEE: "ORB: an efficient alternative to SIFT or SURF", 《2011 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION》 * |
余宏生: "数字图像拼接方法研究进展", 《红外技术》 * |
方贤勇: "图像拼接的改进算法", 《计算机辅助设计与图形学学报》 * |
程争刚: "一种基于无人机位姿信息的航拍图像拼接方法", 《测绘学报》 * |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110569927A (en) * | 2019-09-19 | 2019-12-13 | 浙江大搜车软件技术有限公司 | Method, terminal and computer equipment for scanning and extracting panoramic image of mobile terminal |
CN110796734A (en) * | 2019-10-31 | 2020-02-14 | 中国民航科学技术研究院 | Airport clearance patrol method and device based on high-resolution satellite technology |
CN110796734B (en) * | 2019-10-31 | 2024-01-26 | 中国民航科学技术研究院 | Airport clearance inspection method and device based on high-resolution satellite technology |
CN113228105A (en) * | 2019-12-30 | 2021-08-06 | 商汤国际私人有限公司 | Image processing method and device and electronic equipment |
CN111680703A (en) * | 2020-06-01 | 2020-09-18 | 中国电建集团昆明勘测设计研究院有限公司 | 360-degree construction panorama linkage positioning method based on image feature point detection and matching |
CN111680703B (en) * | 2020-06-01 | 2022-06-03 | 中国电建集团昆明勘测设计研究院有限公司 | 360-degree construction panorama linkage positioning method based on image feature point detection and matching |
CN112308774A (en) * | 2020-09-15 | 2021-02-02 | 北京中科遥数信息技术有限公司 | Unmanned aerial vehicle-based map reconstruction method and system, transmission equipment and storage medium |
CN112184662A (en) * | 2020-09-27 | 2021-01-05 | 成都数之联科技有限公司 | Camera external parameter initial method and system applied to unmanned aerial vehicle image stitching |
CN112184662B (en) * | 2020-09-27 | 2023-12-15 | 成都数之联科技股份有限公司 | Camera external parameter initial method and system applied to unmanned aerial vehicle image stitching |
CN112215304A (en) * | 2020-11-05 | 2021-01-12 | 珠海大横琴科技发展有限公司 | Gray level image matching method and device for geographic image splicing |
CN112907452A (en) * | 2021-04-09 | 2021-06-04 | 长春理工大学 | Optimal suture line searching method for image stitching |
CN113286081A (en) * | 2021-05-18 | 2021-08-20 | 中国民用航空总局第二研究所 | Target identification method, device, equipment and medium for airport panoramic video |
CN117687426A (en) * | 2024-01-31 | 2024-03-12 | 成都航空职业技术学院 | Unmanned aerial vehicle flight control method and system in low-altitude environment |
Also Published As
Publication number | Publication date |
---|---|
CN110033411B (en) | 2021-01-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110033411A (en) | The efficient joining method of highway construction scene panoramic picture based on unmanned plane | |
CN105096386B (en) | A wide range of complicated urban environment geometry map automatic generation method | |
CN104484648B (en) | Robot variable visual angle obstacle detection method based on outline identification | |
CN110285793A (en) | A kind of Vehicular intelligent survey track approach based on Binocular Stereo Vision System | |
CN111209915B (en) | Three-dimensional image synchronous recognition and segmentation method based on deep learning | |
Lu et al. | Geometry-aware satellite-to-ground image synthesis for urban areas | |
CN103236160B (en) | Road network traffic condition monitoring system based on video image processing technology | |
Alonso et al. | Accurate global localization using visual odometry and digital maps on urban environments | |
CN104330074B (en) | Intelligent surveying and mapping platform and realizing method thereof | |
CN110322702A (en) | A kind of Vehicular intelligent speed-measuring method based on Binocular Stereo Vision System | |
CN104156968B (en) | Large-area complex-terrain-region unmanned plane sequence image rapid seamless splicing method | |
CN111126304A (en) | Augmented reality navigation method based on indoor natural scene image deep learning | |
CN106910217A (en) | Vision map method for building up, computing device, computer-readable storage medium and intelligent vehicle | |
CN115439424A (en) | Intelligent detection method for aerial video image of unmanned aerial vehicle | |
Taneja et al. | Registration of spherical panoramic images with cadastral 3d models | |
CN104463778A (en) | Panoramagram generation method | |
CN109685855A (en) | A kind of camera calibration optimization method under road cloud monitor supervision platform | |
CN109520500A (en) | One kind is based on the matched accurate positioning of terminal shooting image and streetscape library acquisition method | |
CN105787876B (en) | One kind being based on the matched panoramic video method for automatically split-jointing of SURF signature tracking | |
CN115388902B (en) | Indoor positioning method and system, AR indoor positioning navigation method and system | |
CN110109465A (en) | A kind of self-aiming vehicle and the map constructing method based on self-aiming vehicle | |
Ma et al. | Boundarynet: extraction and completion of road boundaries with deep learning using mobile laser scanning point clouds and satellite imagery | |
Zhang et al. | Image-only real-time incremental UAV image mosaic for multi-strip flight | |
WO2023123837A1 (en) | Map generation method and apparatus, electronic device, and storage medium | |
US20220214657A1 (en) | Monitoring management and control system based on panoramic big data |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |