CN110378879A - A kind of Bridge Crack detection method - Google Patents
A kind of Bridge Crack detection method Download PDFInfo
- Publication number
- CN110378879A CN110378879A CN201910560732.9A CN201910560732A CN110378879A CN 110378879 A CN110378879 A CN 110378879A CN 201910560732 A CN201910560732 A CN 201910560732A CN 110378879 A CN110378879 A CN 110378879A
- Authority
- CN
- China
- Prior art keywords
- crack
- image
- point
- pixel
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 53
- 238000000034 method Methods 0.000 claims abstract description 45
- 238000012545 processing Methods 0.000 claims abstract description 17
- 239000000284 extract Substances 0.000 claims abstract description 9
- 239000011159 matrix material Substances 0.000 claims description 70
- 230000009466 transformation Effects 0.000 claims description 56
- 238000004422 calculation algorithm Methods 0.000 claims description 35
- 230000011218 segmentation Effects 0.000 claims description 22
- 238000001914 filtration Methods 0.000 claims description 14
- 230000004927 fusion Effects 0.000 claims description 11
- 230000003044 adaptive effect Effects 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 9
- 238000002156 mixing Methods 0.000 claims description 8
- 238000010606 normalization Methods 0.000 claims description 7
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 210000002569 neuron Anatomy 0.000 claims description 6
- 238000010408 sweeping Methods 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 6
- 241000023320 Luma <angiosperm> Species 0.000 claims description 5
- 230000008859 change Effects 0.000 claims description 5
- 238000005336 cracking Methods 0.000 claims description 5
- OSWPMRLSEDHDFF-UHFFFAOYSA-N methyl salicylate Chemical compound COC(=O)C1=CC=CC=C1O OSWPMRLSEDHDFF-UHFFFAOYSA-N 0.000 claims description 5
- 238000013519 translation Methods 0.000 claims description 5
- 238000005516 engineering process Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 4
- 230000009467 reduction Effects 0.000 claims description 4
- 238000001228 spectrum Methods 0.000 claims description 4
- HUTDUHSNJYTCAR-UHFFFAOYSA-N ancymidol Chemical compound C1=CC(OC)=CC=C1C(O)(C=1C=NC=NC=1)C1CC1 HUTDUHSNJYTCAR-UHFFFAOYSA-N 0.000 claims description 3
- 238000013016 damping Methods 0.000 claims description 3
- 230000008555 neuronal activation Effects 0.000 claims description 3
- 238000007430 reference method Methods 0.000 claims description 3
- 230000007704 transition Effects 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 claims description 2
- 238000007689 inspection Methods 0.000 claims description 2
- 230000000007 visual effect Effects 0.000 claims 1
- 238000011160 research Methods 0.000 abstract description 3
- 238000009659 non-destructive testing Methods 0.000 abstract description 2
- 238000004445 quantitative analysis Methods 0.000 abstract 1
- 235000013350 formula milk Nutrition 0.000 description 13
- 238000010586 diagram Methods 0.000 description 12
- 230000000877 morphologic effect Effects 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 5
- 230000006378 damage Effects 0.000 description 5
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 238000012216 screening Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000010276 construction Methods 0.000 description 3
- 201000010099 disease Diseases 0.000 description 3
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 3
- 238000005286 illumination Methods 0.000 description 3
- 238000003709 image segmentation Methods 0.000 description 3
- 238000003763 carbonization Methods 0.000 description 2
- 230000007797 corrosion Effects 0.000 description 2
- 238000005260 corrosion Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000003292 glue Substances 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- BVKZGUZCCUSVTD-UHFFFAOYSA-L Carbonate Chemical compound [O-]C([O-])=O BVKZGUZCCUSVTD-UHFFFAOYSA-L 0.000 description 1
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000013075 data extraction Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000007598 dipping method Methods 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 231100001261 hazardous Toxicity 0.000 description 1
- 230000008595 infiltration Effects 0.000 description 1
- 238000001764 infiltration Methods 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 230000003585 interneuronal effect Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000012528 membrane Substances 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 239000011241 protective layer Substances 0.000 description 1
- 238000000746 purification Methods 0.000 description 1
- 230000006798 recombination Effects 0.000 description 1
- 238000005215 recombination Methods 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 230000003014 reinforcing effect Effects 0.000 description 1
- 230000000246 remedial effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000009394 selective breeding Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization 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
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/35—Determination of transform parameters for the alignment of images, i.e. image registration using statistical methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30132—Masonry; Concrete
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of Bridge Crack detection methods.Since handmarking has certain subjectivity, pavement crack detection accuracy is compared with the Heuristics for depending on expert, and experience lacks objectivity in quantitative analysis.The present invention is as follows: one, acquiring the image for being detected bridge floor one by one, obtain bridge floor image collection.Two, image mosaic.Three, Crack Detection.Four, fracture parameters extract.The present invention completes the automation non-destructive testing of Bridge Crack instead of human eye by Image Acquisition and processing technique, has very important realistic meaning to the research of the Bridge Crack detection technique under complicated landform environment.On the one hand working security is enhanced, on the other hand improves operation mobility and flexibility.The present invention, the fidelity splicing for realizing multiple groups bridge floor image, improve image mosaic precision and efficiency, have established working foundation for subsequent Bridge Crack image detection, also provide a Technical Reference for the image mosaic detection of other field.
Description
Technical field
The invention belongs to technical field of image detection, and in particular to a kind of Bridge Crack detection method.
Background technique
Bridge is built, traffic is improved, effectively promotes the development of Chinese national economy and society.Bridge as connection two across
The important carrier for spending larger location point exposes to the sun and rain and loads in operation by long-term, and generated internal stress also can
Some weak parts are transmitted to along bridge structure, the positional structure surface is caused easily to there is the generation and development in crack.Without
Surface crack with trend is also different to the extent of injury of bridge structure, if the extension trend and structural bearing of surface crack
When face is perpendicular, harm influences maximum.
Through engineering practice and theoretical analysis shows that, most of servicing bridges are all work with cracking, and Bridge Crack band
The potential hazard come should not be underestimated.Once more serious crack occurs in concrete-bridge, outside air and hazardous medium can be very
It is readily permeable to generate carbonate to inside concrete after chemical reaction, cause the basicity environment of wherein reinforcing bar to reduce, surface
Purification membrane is more also easy to produce corrosion after wrecking, in addition, concrete carbonization can also aggravate shrinkage cracking, to the peace of concrete-bridge
Entirely use generation serious harm.As Disease Characters most common in bridge construction, extremely tiny crack (is less than 0.05mm)
Generally structural behaviour is influenced less, to can permit its presence；And compared with large fracture then may load or extraneous physics, chemistry because
Under the action of element, constantly it can generate and extend, be formed through seam, deep seam, even directly affect girder construction indirectly uses the longevity
Life and security performance；If fracture width reaches 0.3mm or more, can directly destroy the globality of structure, cause concrete carbonization,
Protective layer peels off and reinforcement corrosion, forms mechanics discontinuity surface inside bridge, is greatly lowered load carrying capacity of bridge, when serious very
To generation collapse accident, the normal use of Damage Structure.
Therefore, bridge construction damage and the generation and development of its surface crack have much relations.In order to find crack in time,
And adopt remedial measures to eliminate safe hidden trouble, the mode of manual inspection and manual markings is generallyd use, surface crack is by experience
Reviewer's hand dipping abundant and detecting by an unaided eye keeps a record.But the detection mode mobility is poor, risk is big, efficiency
It is low, and since handmarking has certain subjectivity, detection accuracy is compared with the Heuristics for depending on expert, and experience is fixed
Lack objectivity in amount analysis.
Summary of the invention
The purpose of the present invention is to provide a kind of Bridge Crack detection methods.
The specific steps of the present invention are as follows:
Step 1 acquires the image for being detected bridge floor one by one, obtains bridge floor image collection.
Step 2, image mosaic
The pretreatment of 2-1. stitching image
The luma component information of each bridge floor image in bridge floor image collection is extracted and counts, and respectively to each bridge floor image
Luma component information is equalized；Then each bridge floor image is transformed in frequency domain by Fourier transformation, and uses phase
The phase information of normalization crosspower spectrum in related algorithm obtains the translation parameters between image, and completion is overlapped between adjacent image
The pre- estimation in region.
2-2. image registration
Firstly, extracting SIFT feature in overlapping region between each adjacent image；Then pass through adaptive contrast threshold
Method screens SIFT feature, obtains the feature descriptor being made of matching double points；And each neighbor map is calculated using RANSAC algorithm
Projective transformation matrix as between.
2-3. image co-registration
First according to the projective transformation matrix between adjacent image, projective transformation is carried out to corresponding bridge floor image；Then it uses
Be fade-in gradually go out blending algorithm smooth transition is weighted respectively to tri- Color Channel of RGB of each adjacent bridge floor image, obtain bridge floor
Stitching image.
Step 3, Crack Detection
3-1. pre-processes the initial alignment and gray processing for carrying out crack area by Crack Detection.
Gray processing processing is carried out to bridge floor stitching image.Divided by uniform grid bridge floor stitching image is divided into it is multiple
Regional area.The gray scale aggregate-value in each regional area is counted, the intensity profile of all net regions is obtained and extracts bridge floor figure
As average gray value is as gray threshold；Net region of the gray scale aggregate-value lower than gray threshold is filtered out as region of interest
Domain.
3-2. carries out median filtering denoising to each area-of-interest and based on the image enhancement processing of fuzzy set, obtains multiple
Crack gray level image to be split.
3-3. carries out binary segmentation to each crack image to be split by PCNN simplified model, obtains crack segmentation figure
Picture.
3-4. fracture segmented image carries out the removal of remaining isolated noise, obtains crack binary map.
Step 4, fracture parameters extract
The FRACTURE CHARACTERISTICS that 4-1. grow based on seed point to the resulting crack binary map of step 3 connects.
4-2. is identified based on the target classification of rifts of sciagraphy, specific as follows:
First inversion operation is carried out to by step 4-1 treated crack binary map.Using sciagraphy fracture binary map
The projection for carrying out horizontal and vertical direction respectively adds up respectively to each row, column pixel value and sums, and obtains row projection array and column throwing
Shadow array.
According to the actual size feature in crack in the binary map of crack, the length-width ratio of rift portion in the binary map of crack is calculatedΔ x is the floor projection length of rift portion in the binary map of crack, and Δ y is the vertical throwing of rift portion in the binary map of crack
Shadow length.
IfThen determine the crack in the binary map of crack for transverse crack；IfThen determine crack two-value
Crack in figure is longitudinal crack；IfAnd all elements are all larger than 10 in row projection array and column projection array,
Then determine the crack in the binary map of crack for chicken-wire cracking；Otherwise, it is determined that the crack in the binary map of crack is inclined crack；
The extraction of 4-3. deck crack characteristic, to obtain length, width and the area information of practical Bridge Crack.
(1) crack skeleton line is extracted on the crack in the binary map of crack.
Counting the numerical value in the binary map of crack is 1 pixel number as crack elemental area.According to crack binary map
On crack image, extract crack skeleton line.Using the quantity of pixel on the skeleton line of crack as crack length in pixels.
The method for calculating Bridge Crack width is as follows:
(1) one 5 × 5 retrieval template is set, and all slits pixel on fracture skeleton line carries out sweeping, takes
Two farthest crack pixels of distance in the retrieval template, the connecting line of two crack pixels is as template internal fissure part
Strike line.The normal for doing each crack part strike line obtained after sweeping again, seeks its normal angle.
(2) according to the normal angle of crack each on the skeleton line of crack pixel, it is corresponding on crack to seek each crack pixel
Both sides of the edge point (xij,yij) and (x 'ij,y′ij)；The Euclidean distance of crack pixel corresponding both sides of the edge point on crack
As the crack pixel wide d on the location pointij, i.e.,Each crack pixel wide d of gainedij
In maximum value as crack maximum pixel width.Each fracture width dijIn average value as crack mean pixel width.
(3) each pixel-parameters A' in crack is substituted into actual parameter reduction formula respectivelyIn, it is split
Each actual parameter A of seam.Wherein, u is object distance, f is focal length, k is conversion coefficient, and l is CCD sensitive chip long side physical size, L
It is shooting image long side pixel number.The pixel-parameters in crack include that crack length in pixels, crack maximum pixel width, crack are flat
Equal pixel wide and crack elemental area.
Preferably, the method for acquiring image in step 1 is specific as follows:
After 1-1. calculates the internal reference matrix of CCD camera using Zhang Zhengyou plane reference method, diameter is obtained by least square method
To distortion factor.
1-2. disposed on bridge machinery platform by step 1-1 calibration CCD camera, according to preset shooting track into
The Image Acquisition of the complete bridge floor of row.Preset shooting track is S-shaped.
1-3. is according to the obtained CCD camera internal reference matrix of step 1-1 and distortion factor each bridge floor collected to step 1-2
Image carries out image calibration respectively.
Preferably, adaptive contrast threshold method is specific as follows in step 2-2:
(1) characteristic point numerical lower limits N is setmin=200, upper limit Nmax=300, contrast threshold Tc=T0；T0For initial threshold
Value, value are 0.02~0.04.
(2) characteristic point detection is carried out, and Statistical Comparison degree is higher than TcCharacteristic point quantity N.
(3) if Nmin≤N≤Nmax, then contrast is higher than TcCharacteristic point be included in initial matching point set, reject contrast
Lower than threshold value TcCharacteristic point, and be directly entered step (5)；Otherwise, step (4) are executed.
(4) if N < Nmin, then by contrast threshold TcIt is reduced to former numerical valueAnd execute step (3)；If N > Nmax, then
Contrast threshold is increased to 2 times of former numerical value, and executes step (3).
(5) the mistake characteristic point that initial matching point is concentrated is rejected than secondary near neighbor method by arest neighbors, and generates feature description
Symbol.In feature descriptor comprising by pairs of feature point group at multiple matching double points and the distance between each matching double points
And directional information.
Preferably, the process that RANSAC algorithm solves projective transformation matrix is as follows in step 2-2:
(1) original training set S is constructed with each matching double points in feature descriptor.It is respectively matched in statistics original training set S
Euclidean distance between point pair, and by sorting from small to large；
(2) the preceding 85% matching double points building new samples collection S' of sequence obtained by step (1) is taken；
(3) 4 groups of matching double points are randomly selected from new samples collection S' forms an interior point set Si, and calculating matrix model
Interior point set SiHi, enter step (4)；
(4) remaining each matching double points is directed to matrix model H in new samples collection S'iCarry out adaptive test；It examines if it exists
The match point that error is less than error threshold is tested, then interior point set S is added in the matching double points for examining error to be less than threshold valuei, and hold
Row step (5)；Otherwise, give up matrix model Hi, re-execute (3).
(5) if interior point set SiMiddle element number is greater than defined threshold, then it is assumed that reasonable parameter model is obtained, to update
Interior point set S afterwardsiRecalculate matrix model Hi, and use LM algorithmic minimizing cost function；Otherwise, give up the matrix norm
Type Hi, and it re-execute the steps (3).
(6) l step (3) is repeated to (5), and l is maximum number of iterations.Later, interior point set obtained in l iteration is compared
Close Si, with the maximum interior point set S of element numberiAs final interior point set, and the matrix model H for taking it to calculateiAs adjacent
Projective transformation matrix between bridge floor image.
Preferably, specific step is as follows for projective transformation in step 2-3:
(1) according to the transitivity of the projective transformation matrix between adjacent image, made respectively with first bridge floor image of every row
Benchmark image for corresponding row is spliced.Between the transformation matrix H each adjacent bridge floor imageii-1Transmitting transformation is carried out, is obtained
Transmitting transformation matrix H between each bridge floor image and benchmark imagei1；Pass through each transformation matrix H againi1By corresponding bridge floor image
It is respectively mapped in datum plane coordinate system, to complete the image mosaic fusion in horizontal direction between each adjacent image, is formed more
Open the lateral panoramic picture Image of wide viewing anglei。
(2) by first obtained in step (1) lateral panoramic picture Image1Spliced as reference panorama image.
Between the transformation matrix T each lateral panoramic picturejj-1Transmitting transformation is carried out, each lateral panoramic picture Image is obtainediIt is complete with benchmark
Transmitting transformation matrix T between scape imagej1；Pass through each transmitting transformation matrix T againj1Respectively by corresponding lateral panoramic picture point
It is not mapped in datum plane coordinate system, to complete the image mosaic fusion on vertical direction between each adjacent transverse panoramic picture,
Form final bridge floor panoramic picture.
In step 2-3, it is fade-in and gradually goes out in blending algorithm, each merging point pixel value I (x, y) in adjacent image overlapping region
Be fade-in gradually go out weighted formula it is as follows:
Wherein, I1(x,y)、I2(x, y) is respectively correspondence merging point of the two adjacent bridge floor images in overlapping region
Pixel value.d1、d2Gradual change weight factor of the two respectively adjacent bridge floor images in corresponding merging point；Withx1、x2The respectively abscissa on overlapping region two sides boundary；X is the abscissa of corresponding merging point；T is two adjacent
Gray difference threshold of the image overlapping region on corresponding merging point.
Preferably, the step of median filtering denoises in step 3-2 is as follows:
(1) each pixel in area-of-interest is traversed, respectively as target pixel points fij, and respectively execute step (2)
To (4).
(2) respectively to target pixel points fijAll pixels point and target pixel points f in eight neighborhoodijGray value is compared,
And take the wherein the smallest 2 neighborhood territory pixels point of gray scale difference value absolute valueWith
(3) willWithAs the neighbouring expansion of gradation direction of the target pixel points, and with the two sides
To eight neighborhood external expansion level-one, obtaining f respectivelypAnd fqTwo pixels；
(4) target pixel points f is takenij, all pixels point, pixel f in eight neighborhoodpWith pixel fqIntermediate value as target
Pixel fijReplacement values.
Preferably, carrying out binary segmentation to each crack image to be split by PCNN simplified model in step 3-3
The specific method is as follows:
Step1: initialization PCNN model parameter
(1) each normalization pixel (i, j) gray value of crack image to be split is inputted as outside stimulus signal Iij。
(2) damping time constant Δ t=0.2, maximum information entropy H are setmax=0, maximum number of iterations n=30 and chain
Connect strength factor matrix
(3) optimal threshold obtained by two dimension Otsu method, the initial threshold θ as PCNN modelij[0]；Pass through part
Gray variance method determines link strength factor betaij。
Step2: by 1 assignment k.
Step3: iteration is split to crack image to be split using PCNN simplified model, obtains binary segmentation image
Y(k)；Calculate the excited inside U of each neuron in PCNN simplified modelijAnd pulse exports Yij, and by comparing UijWith YijSize
Judge each neuronal activation state.
Step4: information entropy H (k) corresponding to binary segmentation image Y (k) is calculated, if H (k) > Hmax, then H (k) is assigned
It is worth to Hmax, and using binary segmentation image Y (k) as new preferred segmented image Y (s).Otherwise, it is directly entered step5.
Step5: if k < n, after k is increased 1, step Step3 and Step 4 is repeated；Otherwise with current preferred
Segmented image Y (s) is exported as crack segmented image.
Preferably, in step 3-4, remaining isolated noise includes discrete type noise and gathers type and make an uproar spot.
The minimizing technology of discrete type noise is specific as follows:
(1) connected region come is marked off to each edge in the resulting crack segmented image of step 3-3 to be marked.
(2) elemental area of each connected region in best crack segmented image is calculated separately；Elemental area is less than face
The connected region of product threshold value is rejected；Area threshold is 20.
(3) use straightway scanning retrieve: the line segment template length set as 5 pixels, respectively from 0 °, 45 °, 90 ° and
135 ° of four directions are scanned detection to the boundary line for the connected region that step (2) remains, and carry out with operation.If
Straightway scanning retrieval return value of the boundary line of one connected region on 0 °, 45 °, 90 °, 135 ° of directions is 0, then should
Connected region is rejected.
Gather type make an uproar spot minimizing technology it is specific as follows:
(1) the circularity e (x, y) for calculating separately each connected region is as follows:
In formula, C (x, y), A (x, y) are respectively perimeter, the area of corresponding connected region.
If the circularity e (x, y) of a connected region is higher than circularity threshold value Te(x, y) then rejects the connected region,
Te(x, y)=0.3.
(2) the length-width ratio r (x, y) for calculating separately each connected region is as follows:
In formula, L (x, y) and W (x, y) be respectively corresponding connected region in the horizontal and vertical directions projection matrix it is long and
It is wide.
T is setrl(x, y) is length-width ratio lower limit, value 0.1；Trh(x, y) is the length-width ratio upper limit, value 10.If one
The length-width ratio r (x, y) of a connected region meets with lower inequality: Trl(x, y) < r (x, y) < Trh(x, y), then by the connected region
It rejects in domain.
Preferably, crack connection method described in step 4-1 is as follows:
(1) crack gap location information in the binary map of crack is mapped in the resulting crack gray level image of step 3-2.
(2) starting seed point, guidance seed point are chosen respectively in the opposite end in the corresponding two sections of cracks of crack gap.
(3) starting seed point is that first growing point is grown；Next growing point is a upper growing point eight neighborhood
The interior the smallest pixel of gray value, until the growing point grown is overlapped with guidance seed point.
(4) the resulting each growing point of step (3) is mapped in the binary map of crack from the gray level image of crack.
Preferably, after step 4-3 is executed, the sum of crack elemental area in all slits binary map is complete divided by bridge floor
The number of pixels of scape image obtains face crack rate.
The invention has the advantages that:
1, the present invention completes the automation non-destructive testing of Bridge Crack instead of human eye by Image Acquisition and processing technique,
There is very important realistic meaning to the research of the Bridge Crack detection technique under complicated landform environment.On the one hand it enhances and applies
On the other hand work safety improves operation mobility and flexibility.
2, the present invention for tradition based on SIFT feature merging algorithm for images in the problem that operand is larger and precision is insufficient,
In order to more completely and accurately extract bridge floor image FRACTURE CHARACTERISTICS data, a kind of improved picture of large image scale splicing calculation is proposed
Method realizes the fidelity splicing of multiple groups bridge floor image, improves image mosaic precision and efficiency, is subsequent Bridge Crack image detection
Working foundation has been established, has also provided a Technical Reference for the image mosaic detection of other field.
3, the present invention is directed to the unique spatial extension of Bridge Crack and gray scale ga s safety degree, and the improved intermediate value of proposition is filtered
Wave algorithm introduces similar expansion of gradation according to the characteristic of the similar gray scale linear distribution of edge of crack on the basis of former median filtering
The thought in direction can not only more effectively inhibit the interference of bridge floor image recombination noise, improve signal noise ratio (snr) of image, and maintain
The continuity of edge of crack detailed information increases the reliability of target crack fidelity filtering, improves edge of crack feature extraction
The noiseproof feature of algorithm.
4, the crack optimal segmentation image that improved PCNN Crack Detection algorithm proposed by the present invention obtains not only ensure that
Completely and clearly edge contour, and accurate detail textures feature is remained, it effectively eliminates the overwhelming majority and splits
Stitch ambient noise interference adjoint when image segmentation.Compared with traditional images partitioning algorithm, improved pulse coupled neural net
Network algorithm reduces the PCNN parameter that need to manually set, and to the anti-interference under complex background in Bridge Crack image segmentation and divides
Cutting stability has larger promotion, has preferable robustness.
5, the present invention is for discrete type noise in bridge floor image and that gathers that type makes an uproar spot proposition filter out criterion, can be further
The isolated noise region in gray level threshold segmentation treated bianry image is rejected, is thoroughly removed unrelated in bridge floor bianry image
Noise information completely remains the marginal information in target crack, guarantees accuracy and essence that subsequent FRACTURE CHARACTERISTICS data are extracted
Degree.
6, the present invention has carried out multiple groups bridge floor according to collected Bridge Crack image aspects and pixel distribution feature respectively
The algorithm design that image mosaic, large area crack image procossing and FRACTURE CHARACTERISTICS data are extracted, and to wherein key calculate into
Reliability Analysis Research is gone, algorithm novelty with higher and bridge machinery project reference value.
Detailed description of the invention
Fig. 1 is the schematic diagram of Bridge crack detection method of the present invention；
Fig. 2 is bridge floor Image Acquisition track schematic diagram in the present invention；
Fig. 3 is several bridge image mosaic process flow diagrams in the present invention；
Fig. 4 is adaptive contrast threshold calculation flow chart in the present invention；
Fig. 5 is that progressive image splices tactful schematic diagram in the present invention；
Fig. 6 is image mosaic strategy schematic diagram by column in the present invention；
Fig. 7 a, 7b the Weighted Fusion schematic diagram between adjacent image in the present invention；
Fig. 8 is crack area initial alignment schematic diagram in the present invention；
Fig. 9 a, 9b are improved median filtering template schematic diagram in the present invention；
Figure 10 is PCNN simplified model structural schematic diagram in the present invention；
Figure 11 is the improved crack PCNN image flow chart of segmentation algorithm in the present invention；
Figure 12 a, 12b are that seed point growth method crack connection schematic diagram is based in the present invention；
Figure 13 is the local normal direction calculating schematic diagram in the present invention on the skeleton of crack in 5*5 template；
Figure 14 is Bridge calculatingcrackswidth schematic diagram of the present invention.
Specific embodiment
Below in conjunction with attached drawing, the invention will be further described.
As shown in Figure 1, a kind of Bridge Crack detection method, the specific steps are as follows:
Step 1, Image Acquisition
1-1. uses Zhang Zhengyou plane reference method, carries out image sampling to scaling board from different perspectives using CCD camera, leads to
After crossing the internal reference matrix of scaling board X-comers coordinate calculating CCD camera of detection, obtained by least square method radial abnormal
Variable coefficient.
1-2. disposed on bridge machinery platform by step 1-1 calibration CCD camera, according to preset shooting track into
The multiple series of images acquisition of the complete bridge floor of row.Preset shooting track is as shown in Fig. 2, S-shaped.
1-3. is according to the obtained CCD camera internal reference matrix of step 1-1 and distortion factor each bridge floor collected to step 1-2
Image carries out image calibration respectively, to eliminate lens distortion bring distortion effect.
Step 2, image mosaic
The pretreatment of 2-1. stitching image
Estimate the size of panoramic picture in advance first.Resolution ratio sum number measurement value of the size according to image to be spliced, splicing
Inactive area is removed after the completion；Then by extracting and counting the luma component information of all bridge floor images to be spliced, and divide
The other luma component information to each bridge floor image equalizes, to eliminate the influence of uneven illumination bring difference in brightness；Finally
Each bridge floor image is transformed in frequency domain by Fourier transformation, and using the normalization crosspower spectrum in phase related algorithm
Phase information obtains the translation parameters between image, completes the pre- estimation between overlapping region adjacent image with this.
Phase related algorithm is first to be transformed to image to be spliced in frequency domain using Fourier transformation, then pass through normalization mutually
The translation parameters of two image of spectra calculation obtains two-dimensional impulse function: the peak value size reflection of the two-dimensional impulse function
Content relevance between adjacent bridge floor image, value are that 1 two images of expression are identical, indicate entirely different for 0.Bridge figure
Change as brought by being moved between the collected adjacent image of detection device institute there are perspective transform and position, although impulse can be made
The energy of function is dispersed to numerous small leaks from single peak value, but its corresponding translation parameters in peak-peak position can still keep phase
To stabilization.Therefore, the translational movement obtained by phase related algorithm can obtain roughly the overlapping region between image to be spliced, and
The algorithm is insensitive to illumination brightness change, and correlation maximum peak point detected has preferable robustness and stability.
2-2. image registration
Firstly, SIFT feature is extracted in overlapping region between each adjacent image, to reduce a large amount of unnecessary characteristic points
Calculation amount is detected, the detection efficiency of SIFT feature is improved；Then it by adaptive contrast threshold method, will test
SIFT feature quantity controls in a reasonable range, to filter out the stable feature description being made of matching double points
Symbol；And the projective transformation matrix H between each adjacent image is calculated using improved RANSAC algorithm (RANSAC algorithm).
Adaptive contrast threshold method is specific as follows:
Overlapping region between determination two-by-two adjacent bridge floor image (after Δ x, Δ y), is carried out only for the overlapping region
SIFT feature detection.Since wherein the lower characteristic point of contrast is more sensitive to bridge floor ambient noise, therefore set contrast
Threshold value filters out stable feature point set, is denoted as C.
In the prior art, the contrast of each SIFT feature is calculated by difference of Gaussian Taylor expansion, and fixation is set
Contrast threshold retain the characteristic point higher than the contrast threshold as invariant feature point.However above-mentioned contrast threshold Tc
For fixed value, general value is between 0.02 to 0.04.But in the crack image detection of different concrete-bridges, SIFT detection
To candidate feature point set have very big difference, Partial Bridges surface is more smooth bright and clean, and acquired image digital signal is more
Smoothly, scale space factor sigma is smaller, causes the characteristic point detected less, may be unable to satisfy the number of Feature Points Matching instead
Amount demand influences final splicing precision.(the contrast threshold step used in traditional stitching algorithm).The present invention is arranged one
The contrast threshold of variation, to guarantee the SIFT feature number detected control in a reasonable range.It is tested through multiple groups
Verifying, which shows that the characteristic point of Bridge Crack image detection is maintained between 200 to 300, can meet preferable splicing precision.
As shown in figure 4, determining contrast threshold T in the present inventioncMethod, it is specific as follows:
(1) characteristic point numerical lower limits N is setmin=200, upper limit Nmax=300, contrast threshold Tc=T0；T0For initial threshold
Value, value are 0.02~0.04.
(2) characteristic point detection is carried out, and Statistical Comparison degree is higher than TcCharacteristic point quantity N.
(3) if Nmin≤N≤Nmax, then contrast is higher than TcCharacteristic point be included in initial matching point set, reject contrast
Lower than threshold value TcCharacteristic point, and be directly entered step (5)；Otherwise, step (4) are executed.
(4) if N < Nmin, then by contrast threshold TcIt is reduced to former numerical valueAnd execute step (3)；If N > Nmax, then
Contrast threshold is increased to 2 times of former numerical value, and executes step (3).
(5) the mistake characteristic point that initial matching point is concentrated is rejected than secondary near neighbor method by arest neighbors, and generates feature description
Symbol.In feature descriptor comprising by pairs of feature point group at multiple matching double points and the distance between each matching double points
And directional information.
After the Feature Points Matching between adjacent image overlapping region, enough matching double points are filtered out, matching double points are passed through
The transformation matrix between Bridge Crack sequence image is solved, large-scale bridge floor image mosaic is completed with this.To further increase figure
As registration efficiency and precision, RANSAC algorithm is improved.
The process that improved RANSAC algorithm solves projective transformation matrix H is as follows:
(1) original training set S is constructed with each matching double points in feature descriptor.It is respectively matched in statistics original training set S
Euclidean distance between point pair, and by sorting from small to large；
(2) the preceding 85% matching double points building new samples collection S' of sequence obtained by step (1) is taken；
(3) 4 groups of matching double points are randomly selected from new samples collection S' forms an interior point set Si, and calculating matrix model
Interior point set SiHi, enter step (4)；
(4) remaining each matching double points is directed to matrix model H in new samples collection S'iCarry out adaptive test；It examines if it exists
The match point that error is less than error threshold is tested, then interior point set S is added in the matching double points for examining error to be less than threshold valuei, and hold
Row step (5)；Otherwise, give up matrix model Hi, re-execute (3).
(5) if interior point set SiMiddle element number is greater than defined threshold, then it is assumed that reasonable parameter model is obtained, to update
Interior point set S afterwardsiRecalculate matrix model Hi, and use LM algorithmic minimizing cost function；Otherwise, give up the matrix norm
Type Hi, and it re-execute the steps (3).
(6) l step (3) is repeated to (5), and l is maximum number of iterations.Later, interior point set obtained in l iteration is compared
Close Si, with the maximum interior point set S of element numberiAs final interior point set, and the matrix model H for taking it to calculateiAs adjacent
Projective transformation matrix H between bridge floor image.
Improved RANSAC algorithm is by the Euclidean distance between calculating all matching double points and is ranked up screening, not only subtracts
The sample set data for having lacked point pair to be matched improve intra-office point proportion in sample set, and reduce projective transformation square
The iteration of battle array refines number, to improve the matching precision of bridge floor image.According to image characteristic point to the distance between it is smaller,
With the higher characteristic of similarity, calculate herein the Euclidean distance between all characteristic points pair and according to sequence from small to large arrange into
Row screening.After showing overlapped provincial characteristics point to first matching by multiple groups bridge floor image mosaic test result statistics, initial sample
The successful match rate of this collection S then takes preceding 85% characteristic point of its sequence to building new samples collection S' up to 85% or more.Through sample
Notebook data screening, sample set S' include enough matching double points, not only increase intra-office point proportion in sample set, and
Greatly reduce the number of iterations of transformation matrix parameter model H.
2-3. image co-registration
First according to the projective transformation matrix between adjacent image, projective transformation is carried out to corresponding bridge floor image；Then it uses
Be fade-in gradually go out blending algorithm smooth transition is weighted respectively to tri- Color Channel of RGB of each adjacent bridge floor image, obtain bridge floor
Stitching image.
Specific step is as follows for projective transformation:
(1) as shown in figure 5, according to the transitivity of the projective transformation matrix between adjacent image, with first bridge floor of every row
Image is spliced respectively as the benchmark image of corresponding row according to image line splicing strategy.Between each adjacent bridge floor image
Transformation matrix Hii-1Transmitting transformation is carried out, the transmitting transformation matrix H between each bridge floor image and benchmark image is obtainedi1；Pass through again
Each transformation matrix Hi1Corresponding bridge floor image is respectively mapped in datum plane coordinate system, it is each adjacent in horizontal direction to complete
Image mosaic fusion between image, forms the lateral panoramic picture Image of multiple wide viewing anglesi。
Each transfer matrix transformation for mula is as follows:
H21=H21
H31=H32×H21
Hn1=Hnn-1×Hn-1n-2×…×H21
Wherein, Hii-1For the transformation matrix between (i-1)-th bridge floor image and i-th bridge floor image of same a line, value exists
It is calculated in step 2-2；Hi1For the transformation matrix between the 1st bridge floor image of same a line and i-th bridge floor image；N is same
Amount of images in a line.
(2) as shown in fig. 6, first lateral panoramic picture Image obtained in step (1)1As reference panorama image,
According to image column splicing strategy, spliced.Between the transformation matrix T each lateral panoramic picturejj-1Transmitting transformation is carried out, is obtained
Each transverse direction panoramic picture ImageiTransmitting transformation matrix T between reference panorama imagej1；Pass through each transmitting transformation matrix T againj1
Corresponding lateral panoramic picture is respectively mapped in datum plane coordinate system respectively, to complete each adjacent transverse on vertical direction
Image mosaic between panoramic picture merges, and transfer matrix transformation for mula is referring to the description in step (1), shape in image mosaic fusion
At final bridge floor panoramic picture.
It is fade-in in the present embodiment and gradually goes out blending algorithm by improving, referring specifically to following.
After image registration, further to eliminate the interference that bridge floor image mosaic stitches fracture detection processing, usually to phase
Adjacent bridge floor image pixel value is weighted and averaged, as shown in Figure 7b, distance of the pixel to both sides suture in overlapping region
As fusion weight distinguishing rule.
But since image capture position changes, bridge surface reflection is likely to result in respective pixel in overlapping region
There are hopping phenomenons for point gray value, to eliminate the influence that it generates blending image, are fade-in in tradition and gradually go out Weighted Fusion calculating
One threshold value t of middle introducing.Lap target pixel points are calculated in the corresponding gray scale difference value of two width original images, if the difference is small
In threshold value, illustrate the pixel in former bridge floor image and not shown notable difference, can directly take its weighted average as should
Point pixel value；Conversely, illustrating that image to be spliced, there are light and shade mutation, should take its smooth preceding weight larger under the pixel position
Pixel value as the fusion pixel values.
It is fade-in and gradually goes out in blending algorithm, being fade-in for each fusion pixel values I (x, y) gradually goes out to weight in adjacent image overlapping region
Formula is as follows:
Wherein, I1(x,y)、I2(x, y) is respectively correspondence merging point of the two adjacent bridge floor images in overlapping region
Pixel value, as shown in Figure 7a.d1、d2Gradual change weight factor of the two respectively adjacent bridge floor images in corresponding merging point；Such as figure
Shown in 7b,Withx1、x2The respectively abscissa on overlapping region two sides boundary；X is corresponding merging point
Abscissa；T is gray difference threshold of the two adjacent image overlapping regions on corresponding merging point.
Step 3, Crack Detection
3-1. pre-processes the initial alignment and gray processing for carrying out crack area by Crack Detection.
Gray processing processing is carried out to bridge floor stitching image.Divided by uniform grid bridge floor stitching image is divided into it is multiple
Regional area.Since Bridge object crack gray value will be lower than local background's gray average, therefore count the ash in each regional area
Aggregate-value is spent, the intensity profile of all net regions is obtained and extracts bridge floor image averaging gray value as gray threshold；Screening
Gray scale aggregate-value carries out the initial alignment mark of crack target area lower than the net region of gray threshold as area-of-interest out
Note, as shown in Figure 8.
3-2. is denoised to the median filtering that each area-of-interest improves and the image enhancement processing based on fuzzy set, obtains
The crack gray level image to be split to multiple.
Although crack gray level image lacks color contrast abundant, there is unique spatial extension and gray scale can distinguish
Property.By being improved to median filtering, according to uniform one of linear character and edge pixel gray scale under the local space of crack
Cause property, introduces similar expansion of gradation direction, as illustrated in fig. 9, by the adjacent pixels on its direction on the basis of former median filtering
Point is included in Filtering Template, reduces neighbouring noise under former Filtering Template to a certain extent and interferes the filtering of center pixel.
The step of median filtering denoises is as follows:
(1) each pixel in area-of-interest is traversed, respectively as target pixel points fij, and respectively execute step (2)
To (4).
(2) respectively to target pixel points fijAll pixels point and target pixel points f in eight neighborhoodijGray value is compared,
And take the wherein the smallest 2 neighborhood territory pixels point of gray scale difference value absolute valueWith
(3) willWithAs the neighbouring expansion of gradation direction of the target pixel points, as shown in figure 9b, and
With the two directions respectively to eight neighborhood external expansion level-one, f is obtainedpAnd fqTwo pixels；
(4) to target pixel points fij, all pixels point, pixel f in eight neighborhoodpWith pixel fqThe array of composition carries out
Gray value sequence, takes the intermediate value of the array as target pixel points fijReplacement values.
3-3. carries out two-value point to each crack image to be split by PCNN (Pulse Coupled Neural Network) simplified model
It cuts；
The present invention is to improve crack image segmentation efficiency, and PCNN simplified model is used under the premise of guaranteeing detection accuracy,
As shown in Figure 10, partial parameters influence is eliminated.And according to the space under complicated bridge floor background between target crack and neighborhood territory pixel
Location information is proposed by the optimal initial threshold value in two dimension Otsu algorithm automatic calculation gray level image, to guarantee model for the first time
Correct nerve impulse is generated when iteration, improves the Searching efficiency of best iterative segmentation effect；In conjunction with neuron coupled characteristic and
Image space gray-scale watermark obtains interneuronal link strength by solving Bridge Crack image local gray scale mean square deviation
Coefficient, it is same to capture the neuron within the scope of local gray level as stiffness of coupling accurate between neuron in regional area
Pace pulse.
The specific method for carrying out binary segmentation to each crack image to be split by PCNN simplified model is as shown in figure 11,
It is described in detail below:
Step1: initialization PCNN model parameter
(1) each normalization pixel (i, j) gray value of crack image to be split is inputted as outside stimulus signal Iij,
That is Fij(n)=Iij。
(2) damping time constant Δ t=0.2, maximum information entropy H are setmax=0, maximum number of iterations n=30 and chain
Connect strength factor matrix
(3) optimal threshold obtained by two dimension Otsu method, the initial threshold θ as PCNN modelij[0]；Pass through part
Gray variance method determines link strength factor betaij。
Step2: by 1 assignment k.
Step3: iteration is split to crack image to be split using PCNN simplified model, obtains binary segmentation image
Y (k) is divided obtained binary segmentation image every time and is all different；Calculate the excited inside of each neuron in PCNN simplified model
UijAnd pulse exports Yij, and by comparing UijWith YijSize judges each neuronal activation state.
Step4: information entropy H (k) corresponding to binary segmentation image Y (k) is calculated, if H (k) > Hmax, then H (k) is assigned
It is worth to Hmax, and using binary segmentation image Y (k) as new preferred segmented image Y (s).Otherwise, it is directly entered step5.
Step5: if k < n, after k is increased 1, step Step3 and Step 4 is repeated；Otherwise, maximum informational entropy is taken
HmaxCorresponding the number of iterations is as best the number of iterations kH, and using current preferred segmented image Y (s) as best crack
Segmented image output.
3-4. carries out the removal of remaining isolated noise to best crack segmented image, obtains crack binary map.
Due to bridge surface image can with different degrees of interference information, as own face stain, shooting when illumination
Uneven bring noise spot etc. causes to make an uproar in the bridge floor bianry image after edge of crack dividing processing there are still isolated on a small quantity
Sound is classified as discrete type noise and gathers type making an uproar two class of spot according to noise profile characteristic.
The distribution of discrete type noise is more sparse in bridge floor image, and pixel quantity is smaller, often show as it is small make an uproar spot or
The characteristics of short line.Therefore, according to the morphological feature of glue into concrete beam cracks, binaryzation edge is carried out using long line segment template
With operation determines the length characteristic of each connected region edge line, each connection surrounded with area threshold to binaryzation edge
Region area is respectively calculated and differentiates the area features of isolated area.Steps are as follows for specific algorithm:
(1) each edge marks off the connected region progress come in best crack segmented image resulting to step 3-3
Label.
(2) elemental area of each connected region in best crack segmented image is calculated separately；Elemental area is less than face
The connected region of product threshold value is considered as isolated noise region, is rejected；Area threshold is 20.
(3) use straightway scanning retrieve: the line segment template length set as 5 pixels, respectively from 0 °, 45 °, 90 ° and
135 ° of four directions are scanned detection to the boundary line for the connected region that step (2) remains, and carry out with operation.If
Straightway scanning retrieval return value of the boundary line of one connected region on 0 °, 45 °, 90 °, 135 ° of directions is 0 (i.e. four
The straight line of continuous 5 pixels is not present on direction), then the connected region is considered as isolated noise region, is rejected.
Gather type in bridge floor image and makes an uproar spot often by the formation such as the dirty pool in surface, hollow and shadow occlusion, inner part
Cloth is more closely concentrated, and area is larger, at bulk, according to its features of shape, is divided into rule and is made an uproar spot and spot of irregularly making an uproar.
Removal gather type make an uproar spot method it is specific as follows:
(1) according to the lesser feature of crack circularity, standard is filtered out as regular spot of making an uproar using circularity.
The circularity e (x, y) for calculating separately each connected region is as follows:
In formula, C (x, y), A (x, y) are respectively perimeter, the area of corresponding connected region.
If the circularity e (x, y) of a connected region is higher than circularity threshold value TeThe connected region is then considered as by (x, y)
Rule is made an uproar spot, is rejected, Te(x, y)=0.3.
(2) according to the morphological feature of crack elongate curved, mark is filtered out as spot region of making an uproar using certain length and width ratio
It is quasi-.
The length-width ratio r (x, y) for calculating separately each connected region is as follows:
In formula, L (x, y) and W (x, y) be respectively corresponding connected region in the horizontal and vertical directions projection matrix it is long and
It is wide.
In view of transverse crack length and width ratio is high, longitudinal crack is then extremely low, therefore, T is arrangedrl(x, y) is under length-width ratio
Limit, value 0.1；Trh(x, y) is the length-width ratio upper limit, value 10.If the length-width ratio r (x, y) of connected region meet with
Lower inequality: Trl(x, y) < r (x, y) < TrhThe connected region is then considered as small area and irregularly made an uproar spot region by (x, y), into
Row is rejected.
Step 4, data extraction module
The FRACTURE CHARACTERISTICS that 4-1. grow based on seed point to the resulting crack binary map of step 3 connects, and to complete
Crack image carry out morphologic thinning and deburring processing, retain the most key crack body feature.
The consistency in present invention edge tangent line direction according to locating for front and back fracture two-port, proposes using based on seed point
The crack join algorithm of growth holds in the mouth it by the way that the starting seed point at adjacent crack port and guidance seed point is arranged
Connect fitting.Starting seed point is first seed point that growth starts, and indicates the initial position of region growing；Guidance seed point be
The seed point of guidance starting seed point linking growth, indicates the termination area that fitting extends.But due to from starting seed point to drawing
The fitting growth of seed point is led depending on the intensity profile on extending direction, often will appear growth and deviate.It is necessary to set
Threshold value deviates if there is growth, only needs its match point to guide within the scope of threshold distance of the seed point as the center of circle, be equally considered as
Successful connection is grown, otherwise, selected seed point is needed again to be fitted.
Its crack connection method is as follows:
(1) since the bianry image after edge is divided no longer contains crack neighborhood grayscale information, it is raw to be unable to satisfy region
Long requirement.By artificial selection or utilize the resulting crack of branchpoints function obtaining step 3 in MATBAL software
Gap position in crack in binary map.Gained breaking part location information is mapped in the resulting crack gray level image of step 3-2,
Carry out seed point growth.
(2) starting seed point, guidance seed point are chosen respectively in the opposite end in the corresponding two sections of cracks of crack gap, from
And determine the essential adaptor direction of growth in crack at this, so that the extending direction better authenticity.
(3) in order to more precisely be fitted due to the connecting line that contrast is low and accidentally divides, need to formulate seed point growth
Rule.Since crack pixel value is lower in gray level image.It is that first growing point is grown to originate seed point；Next life
Long point is the smallest pixel of gray value in a upper growing point eight neighborhood, the growing point until growing and guidance seed point weight
It closes.
(4) the resulting each growing point of step (3) is mapped in the binary map of crack from the gray level image of crack, completes crack
The complete connection of breaking part.
The present embodiment grows legend as presented example, as shown in the left side Figure 12 a, grid matrix representative using crack seed point
Image slices vegetarian refreshments, numerical value indicate gray value, grey grid represents at two disconnected pixel at the port of crack, has black
P the and q point of circle respectively indicates selected starting seed point and guidance seed point.Previous growing point eight neighborhood direction is as schemed
Shown in 12b, judgement crack connecting line direction is the second extending direction, then provides that the second extending direction is principal direction, and press the inverse time
Then, setting first and third extending direction are respectively pair one and secondary two directions to divider, and with the priority of its neighborhood direction of growth
Successively reduce.Each growing point p in the direction of growth is constantly determined with thisi, indicated with heavy black line frame, and its coordinate points is mapped
Into crack bianry image, as shown in the right side Figure 12, crack connecting line at this can be used as.
4-2. is identified based on the target classification of rifts of sciagraphy, according to its type of deck crack morphological feature automatic discrimination；
First inversion operation is carried out to by step 4-1 treated crack binary map.Further according to crack in the binary map of crack
The characteristics of edge line pixel is 1, and bridge floor background pixel is 0 is carried out horizontal and vertical respectively using sciagraphy fracture binary map
The projection in direction, to each row, column pixel value respectively add up summation and will gained and value deposit array in, obtain row projection array and
Column projection array.According to the differentiation for carrying out types of fractures the characteristics of two groups of obtained projection arrays, by glue into concrete beam cracks point
For lateral, longitudinal, oblique and chicken-wire cracking.In common deck crack classification, the appearance features in different type crack are more
Obviously, laterally, longitudinal direction and inclined crack all show as preferable linear character, infiltration and development direction is more clear, and netted splits
Seam then shows as poor directionality.By all kinds of cracks morphological feature different in bianry image, as target crack point
The foundation of class identification.
Due to the continuity of edge of crack line, so that the array that horizontal and vertical direction projection comes out also has continuity,
Thus x (i), y (j) are respectively i-th of the numerical value and column projection j-th of numerical value of array of the row projection array of crack binary map.Its
Expression formula is as follows:
Wherein, f (i, j) is the pixel value that crack binary map is the i-th row jth column；M, N is respectively the length and width of crack binary map
Size.
The specific method for differentiating crack binary map types of fractures is as follows:
According to the actual size feature in crack in the binary map of crack, the length-width ratio of rift portion in the binary map of crack is calculatedΔ x is the floor projection length of rift portion in the binary map of crack, and Δ y is the vertical throwing of rift portion in the binary map of crack
Shadow length.
IfThen determine the crack in the binary map of crack for transverse crack；IfThen determine crack two-value
Crack in figure is longitudinal crack；IfAnd all elements are all larger than 10 in row projection array and column projection array,
Then illustrate that the edge of crack many places overlapping all occurs in the projection of both direction, determines that the crack in the binary map of crack is netted
Crack；Otherwise, it is determined that the crack in the binary map of crack is inclined crack；
The extraction of 4-3. deck crack characteristic, to obtain length, width and the area information of practical Bridge Crack.
Bridge Crack detection is final or in order to obtain the Disease Characters information in target crack, and controls in this, as observation
The important evidence administered with maintenance.Therefore, after the positioning of target crack area and morphological feature identification classification, it is also necessary to fracture
Morphological feature parameter carries out quantitative data calculating.The morphological feature parameter of Bridge Crack mainly include area, length, width and
Maximum width, wherein fracture width information is to assess the key index of Bridge Crack degree of disease.
(1) crack skeleton line is extracted on the crack in the binary map of crack.
Crack skeleton line of the crack image after edge of crack segmentation and the removal of burr branch in fracture binary map,
The fissured central line being made of most representative single pixel.Due to the measurement accuracy of bridge machinery fracture length and area
It is required that not high, in order to improve computational efficiency, the quantity of the directly pixel on statistics crack skeleton line is long as crack pixel
Degree；Counting the numerical value in the binary map of crack is 1 pixel number as crack elemental area.
Due to the crack skeleton after edge thinning and burr removal, the true trend and hair of deck crack can reflect
Open up situation.By retrieving crack Skeleton pixel point, the vertical direction of each position point is obtained, edge of crack bianry image is re-mapped
In, Bridge Crack width information is calculated, as shown in figure 13, it is as follows specifically to solve process:
(1) as shown in figure 14, one 5 × 5 retrieval template, and all slits pixel on fracture skeleton line are set
Sweeping is carried out, two crack pixels taking distance in the retrieval template farthest (i.e. numerical value be 1 pixel), two slit images
The connecting line of vegetarian refreshments is as template internal fissure part strike line.The normal of each crack part strike line obtained after sweeping is done again,
Seek its normal angle.
(2) according to the normal angle of crack each on the skeleton line of crack pixel, it is corresponding on crack to seek each crack pixel
Both sides of the edge point (xij,yij) and (x 'ij,y′ij)；Pixel corresponding both sides of the edge point on crack in crack is the crack
The normal of pixel and the intersection point of crack both sides of the edge line.The Euclidean of crack pixel corresponding both sides of the edge point on crack away from
From as the crack pixel wide d on the location pointij, i.e.,Each crack pixel wide of gained
dijIn maximum value as crack maximum pixel width.Each fracture width dijIn average value as crack mean pixel width.
(3) each pixel-parameters A' in crack is substituted into actual parameter reduction formula respectivelyIn, it is split
Each actual parameter A of seam.Wherein, u is object distance, f is focal length, k is conversion coefficient, and l is CCD sensitive chip long side physical size, L
It is shooting image long side pixel number.The pixel-parameters in crack include that crack length in pixels, crack maximum pixel width, crack are flat
Equal pixel wide and crack elemental area.By the sum of crack elemental area in all slits binary map divided by bridge floor panoramic picture
Number of pixels, obtain face crack rate.
The derivation process of actual parameter reduction formula is as follows:
Image distance can be obtained by lens imaging formula:
Then optical magnification:
The imaging size in target crack are as follows: A'=MA
Wherein, it be image distance, f be focal length, A be crack target actual size, A ' is object pixel size that u, which is object distance, v, but
Since A is as unit of mm, A ' is as unit of pixel, so also needing to carry out unit conversion, wherein k value is the system that converts
Number, l are that CCD sensitive chip long side physical size, L are shooting image long side pixel numbers.
Its practical physical message can be obtained from there through the data information for extracting image crack:To realize the detection of location information and dimension information of the Bridge Crack in true detection faces
It calculates.
Testing number can be carried out by the sensitive chip size and pixel resolution and bridge floor shooting distance of shooting camera
According to conversion, and need to carry out the adjustment of equipment according to the precision of bridge machinery engineering in advance.
Claims (10)
1. a kind of Bridge Crack detection method, it is characterised in that: step 1 acquires the image for being detected bridge floor one by one, obtains bridge floor
Image collection；
Step 2, image mosaic
The pretreatment of 2-1. stitching image
The luma component information of each bridge floor image in bridge floor image collection is extracted and counts, and the brightness to each bridge floor image respectively
Component information is equalized；Then each bridge floor image is transformed in frequency domain by Fourier transformation, and related using phase
The phase information of normalization crosspower spectrum in algorithm obtains the translation parameters between image, completes between overlapping region adjacent image
Pre- estimation；
2-2. image registration
Firstly, extracting SIFT feature in overlapping region between each adjacent image；Then it is sieved by adaptive contrast threshold method
SIFT feature is selected, the feature descriptor being made of matching double points is obtained；And using between each adjacent image of RANSAC algorithm calculating
Projective transformation matrix；
2-3. image co-registration
First according to the projective transformation matrix between adjacent image, projective transformation is carried out to corresponding bridge floor image；Then it uses and is fade-in
Gradually go out blending algorithm and smooth transition is weighted respectively to tri- Color Channel of RGB of each adjacent bridge floor image, obtains bridge floor splicing
Image；
Step 3, Crack Detection
3-1. pre-processes the initial alignment and gray processing for carrying out crack area by Crack Detection；
Gray processing processing is carried out to bridge floor stitching image；It is divided by uniform grid and bridge floor stitching image is divided into multiple parts
Region；The gray scale aggregate-value in each regional area is counted, the intensity profile of all net regions is obtained and extracts bridge floor image is flat
Equal gray value is as gray threshold；Net region of the gray scale aggregate-value lower than gray threshold is filtered out as area-of-interest；
3-2. carries out median filtering denoising and based on the image enhancement processing of fuzzy set to each area-of-interest, obtains multiple and waits for point
The crack gray level image cut；
3-3. carries out binary segmentation to each crack image to be split by PCNN simplified model, obtains crack segmented image；
3-4. fracture segmented image carries out the removal of remaining isolated noise, obtains crack binary map；
Step 4, fracture parameters extract
The FRACTURE CHARACTERISTICS that 4-1. grow based on seed point to the resulting crack binary map of step 3 connects；
4-2. is identified based on the target classification of rifts of sciagraphy, specific as follows:
First inversion operation is carried out to by step 4-1 treated crack binary map；Distinguished using sciagraphy fracture binary map
The projection for carrying out horizontal and vertical direction adds up respectively to each row, column pixel value and sums, and obtains row projection array and column project number
Group；
According to the actual size feature in crack in the binary map of crack, the length-width ratio of rift portion in the binary map of crack is calculatedΔx
For the floor projection length of rift portion in the binary map of crack, Δ y is the vertical projection length of rift portion in the binary map of crack；
IfThen determine the crack in the binary map of crack for transverse crack；IfThen determine in the binary map of crack
Crack is longitudinal crack；IfAnd all elements are all larger than 10 in row projection array and column projection array, then determine
Crack in the binary map of crack is chicken-wire cracking；Otherwise, it is determined that the crack in the binary map of crack is inclined crack；
The extraction of 4-3. deck crack characteristic, to obtain length, width and the area information of practical Bridge Crack；
(1) crack skeleton line is extracted on the crack in the binary map of crack；
Counting the numerical value in the binary map of crack is 1 pixel number as crack elemental area；According in the binary map of crack
Crack image extracts crack skeleton line；Using the quantity of pixel on the skeleton line of crack as crack length in pixels；
The method for calculating Bridge Crack width is as follows:
(1) one 5 × 5 retrieval template is set, and all slits pixel on fracture skeleton line carries out sweeping, takes the inspection
The connecting line of two farthest crack pixels of distance in rope template, two crack pixels is locally moved towards as template internal fissure
Line；The normal for doing each crack part strike line obtained after sweeping again, seeks its normal angle；
(2) according to the normal angle of crack each on the skeleton line of crack pixel, each crack pixel is sought corresponding two on crack
Side edge point (xij,yij) and (x 'ij,y′ij)；The Euclidean distance conduct of crack pixel corresponding both sides of the edge point on crack
Crack pixel wide d on the location pointij, i.e.,Each crack pixel wide d of gainedijIn
Maximum value is as crack maximum pixel width；Each fracture width dijIn average value as crack mean pixel width；
(3) each pixel-parameters A' in crack is substituted into actual parameter reduction formula respectivelyIn, obtain crack
Each actual parameter A；Wherein, u is object distance, f is focal length, k is conversion coefficient, and l is CCD sensitive chip long side physical size, L is to clap
Take the photograph image long side pixel number；The pixel-parameters in crack include that crack length in pixels, crack maximum pixel width, crack are averaged picture
Plain width and crack elemental area.
2. a kind of Bridge Crack detection method according to claim 1, it is characterised in that: acquire the side of image in step 1
Method is specific as follows:
After 1-1. calculates the internal reference matrix of CCD camera using Zhang Zhengyou plane reference method, obtained by least square method radial abnormal
Variable coefficient；
1-2. disposes the CCD camera by step 1-1 calibration on bridge machinery platform, has been carried out according to preset shooting track
The Image Acquisition of whole bridge floor；Preset shooting track is S-shaped；
1-3. is according to the obtained CCD camera internal reference matrix of step 1-1 and distortion factor each bridge floor image collected to step 1-2
Image calibration is carried out respectively.
3. a kind of Bridge Crack detection method according to claim 1, it is characterised in that: adaptive to compare in step 2-2
It is specific as follows to spend threshold method:
(1) characteristic point numerical lower limits N is setmin=200, upper limit Nmax=300, contrast threshold Tc=T0；T0For initial threshold,
Value is 0.02~0.04；
(2) characteristic point detection is carried out, and Statistical Comparison degree is higher than TcCharacteristic point quantity N；
(3) if Nmin≤N≤Nmax, then contrast is higher than TcCharacteristic point be included in initial matching point set, reject contrast and be lower than threshold
Value TcCharacteristic point, and be directly entered step (5)；Otherwise, step (4) are executed；
(4) if N < Nmin, then by contrast threshold TcIt is reduced to former numerical valueAnd execute step (3)；If N > Nmax, then will be right
Than 2 times that degree threshold value is increased to former numerical value, and execute step (3)；
(5) the mistake characteristic point that initial matching point is concentrated is rejected than secondary near neighbor method by arest neighbors, and generates feature descriptor；It is special
Levy in descriptor comprising by pairs of feature point group at multiple matching double points and the distance between each matching double points and direction
Information.
4. a kind of Bridge Crack detection method according to claim 1, it is characterised in that: in step 2-2, RANSAC algorithm
The process for solving projective transformation matrix is as follows:
(1) original training set S is constructed with each matching double points in feature descriptor；Count each matching double points in original training set S
Between Euclidean distance, and by sorting from small to large；
(2) the preceding 85% matching double points building new samples collection S' of sequence obtained by step (1) is taken；
(3) 4 groups of matching double points are randomly selected from new samples collection S' forms an interior point set Si, and point in calculating matrix model
Set SiHi, enter step (4)；
(4) remaining each matching double points is directed to matrix model H in new samples collection S'iCarry out adaptive test；Error is examined if it exists
Less than the match point of error threshold, then interior point set S is added in the matching double points for examining error to be less than threshold valuei, and execute step
(5)；Otherwise, give up matrix model Hi, re-execute (3)；
(5) if interior point set SiMiddle element number is greater than defined threshold, then it is assumed that reasonable parameter model is obtained, to updated
Interior point set SiRecalculate matrix model Hi, and use LM algorithmic minimizing cost function；Otherwise, give up matrix model Hi,
And it re-execute the steps (3)；
(6) l step (3) is repeated to (5), and l is maximum number of iterations；Later, interior point set S obtained in l iteration is comparedi,
With the maximum interior point set S of element numberiAs final interior point set, and the matrix model H for taking it to calculateiAs adjacent bridge floor
Projective transformation matrix between image.
5. a kind of Bridge Crack detection method according to claim 1, it is characterised in that: in step 2-3, projective transformation
Specific step is as follows:
(1) according to the transitivity of the projective transformation matrix between adjacent image, using first bridge floor image of every row as right
The benchmark image that should be gone is spliced；Between the transformation matrix H each adjacent bridge floor imageii-1Transmitting transformation is carried out, each bridge is obtained
Transmitting transformation matrix H between face image and benchmark imagei1；Pass through each transformation matrix H againi1Corresponding bridge floor image is distinguished
It is mapped in datum plane coordinate system, to complete the image mosaic fusion in horizontal direction between each adjacent image, forms multiple width
The lateral panoramic picture Image at visual anglei；
(2) by first obtained in step (1) lateral panoramic picture Image1Spliced as reference panorama image；To each
Transformation matrix T between lateral panoramic picturejj-1Transmitting transformation is carried out, each lateral panoramic picture Image is obtainediWith reference panorama figure
Transmitting transformation matrix T as betweenj1；Pass through each transmitting transformation matrix T againj1Corresponding lateral panoramic picture is reflected respectively respectively
It is mapped in datum plane coordinate system, to complete the image mosaic fusion on vertical direction between each adjacent transverse panoramic picture, is formed
Final bridge floor panoramic picture；
In step 2-3, it is fade-in and gradually goes out in blending algorithm, each merging point pixel value I (x, y) is fade-in in adjacent image overlapping region
It is as follows gradually to go out weighted formula:
Wherein, I1(x,y)、I2(x, y) is respectively the pixel of correspondence merging point of the two adjacent bridge floor images in overlapping region
Value；d1、d2Gradual change weight factor of the two respectively adjacent bridge floor images in corresponding merging point；Withx1、x2The respectively abscissa on overlapping region two sides boundary；X is the abscissa of corresponding merging point；T is two adjacent
Gray difference threshold of the image overlapping region on corresponding merging point.
6. a kind of Bridge Crack detection method according to claim 1, it is characterised in that: median filtering is gone in step 3-2
The step of making an uproar is as follows:
(1) each pixel in area-of-interest is traversed, respectively as target pixel points fij, and respectively execute step (2) extremely
(4)；
(2) respectively to target pixel points fijAll pixels point and target pixel points f in eight neighborhoodijGray value is compared, and is taken
The wherein the smallest 2 neighborhood territory pixels point of gray scale difference value absolute valueWith
(3) willWithAs the neighbouring expansion of gradation direction of the target pixel points, and with the two directions point
Not to eight neighborhood external expansion level-one, f is obtainedpAnd fqTwo pixels；
(4) target pixel points f is takenij, all pixels point, pixel f in eight neighborhoodpWith pixel fqIntermediate value as object pixel
Point fijReplacement values.
7. a kind of Bridge Crack detection method according to claim 1, it is characterised in that: in step 3-3, pass through PCNN letter
Changing model, the specific method is as follows to each crack image progress binary segmentation to be split:
Step1: initialization PCNN model parameter
(1) each normalization pixel (i, j) gray value of crack image to be split is inputted as outside stimulus signal Iij；
(2) damping time constant Δ t=0.2, maximum information entropy H are setmax=0, maximum number of iterations n=30 and link are strong
Spend coefficient matrix
(3) optimal threshold obtained by two dimension Otsu method, the initial threshold θ as PCNN modelij[0]；Pass through local gray level
Variance method determines link strength factor betaij；
Step2: by 1 assignment k；
Step3: iteration is split to crack image to be split using PCNN simplified model, obtains binary segmentation image Y
(k)；Calculate the excited inside U of each neuron in PCNN simplified modelijAnd pulse exports Yij, and by comparing UijWith YijSize
Judge each neuronal activation state；
Step4: information entropy H (k) corresponding to binary segmentation image Y (k) is calculated, if H (k) > Hmax, then H (k) is assigned to
Hmax, and using binary segmentation image Y (k) as new preferred segmented image Y (s)；Otherwise, it is directly entered step5；
Step5: if k < n, after k is increased 1, step Step3 and Step4 are repeated；Otherwise with current preferred segmentation figure
As Y (s) is exported as crack segmented image.
8. a kind of Bridge Crack detection method according to claim 1, it is characterised in that: in step 3-4, remnants are isolated to make an uproar
Sound includes discrete type noise and gathers type and make an uproar spot；
The minimizing technology of discrete type noise is specific as follows:
(1) connected region come is marked off to each edge in the resulting crack segmented image of step 3-3 to be marked；
(2) elemental area of each connected region in best crack segmented image is calculated separately；Elemental area is less than area threshold
The connected region of value is rejected；Area threshold is 20；
(3) use straightway scanning to retrieve: the line segment template length set is 5 pixel, respectively from 0 °, 45 °, 90 ° and 135 °
Four direction is scanned detection to the boundary line for the connected region that step (2) remains, and carries out with operation；If one
Straightway scanning retrieval return value of the boundary line of connected region on 0 °, 45 °, 90 °, 135 ° of directions is 0, then by the connection
It rejects in region；
Gather type make an uproar spot minimizing technology it is specific as follows:
(1) the circularity e (x, y) for calculating separately each connected region is as follows:
In formula, C (x, y), A (x, y) are respectively perimeter, the area of corresponding connected region；
If the circularity e (x, y) of a connected region is higher than circularity threshold value Te(x, y) then rejects the connected region, Te(x,
Y)=0.3；
(2) the length-width ratio r (x, y) for calculating separately each connected region is as follows:
In formula, L (x, y) and W (x, y) are respectively that projection matrix is long and wide in the horizontal and vertical directions for corresponding connected region；
T is setrl(x, y) is length-width ratio lower limit, value 0.1；Trh(x, y) is the length-width ratio upper limit, value 10；If a connection
The length-width ratio r (x, y) in region meets with lower inequality: Trl(x, y) < r (x, y) < Trh(x, y) then picks the connected region
It removes.
9. a kind of Bridge Crack detection method according to claim 1, it is characterised in that: crack described in step 4-1
Connection method is as follows:
(1) crack gap location information in the binary map of crack is mapped in the resulting crack gray level image of step 3-2；
(2) starting seed point, guidance seed point are chosen respectively in the opposite end in the corresponding two sections of cracks of crack gap；
(3) starting seed point is that first growing point is grown；Next growing point is ash in a upper growing point eight neighborhood
The smallest pixel of angle value, until the growing point grown is overlapped with guidance seed point；
(4) the resulting each growing point of step (3) is mapped in the binary map of crack from the gray level image of crack.
10. a kind of Bridge Crack detection method according to claim 1, it is characterised in that: after step 4-3 is executed, by institute
There is the sum of crack elemental area in the binary map of crack divided by the number of pixels of bridge floor panoramic picture, obtains face crack rate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910560732.9A CN110378879B (en) | 2019-06-26 | 2019-06-26 | Bridge crack detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910560732.9A CN110378879B (en) | 2019-06-26 | 2019-06-26 | Bridge crack detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110378879A true CN110378879A (en) | 2019-10-25 |
| CN110378879B CN110378879B (en) | 2021-03-02 |
Family
ID=68249462
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910560732.9A Active CN110378879B (en) | 2019-06-26 | 2019-06-26 | Bridge crack detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110378879B (en) |
Cited By (37)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110926342A (en) * | 2019-11-27 | 2020-03-27 | 北京工业大学 | Crack width measuring method and device |
| CN110969576A (en) * | 2019-11-13 | 2020-04-07 | 同济大学 | Highway pavement image splicing method based on roadside PTZ camera |
| CN110991466A (en) * | 2019-11-21 | 2020-04-10 | 同济大学 | Highway road surface condition detecting system based on novel vision sensing equipment |
| CN111008956A (en) * | 2019-11-13 | 2020-04-14 | 武汉工程大学 | Beam bottom crack detection method, system, device and medium based on image processing |
| CN111080514A (en) * | 2019-11-07 | 2020-04-28 | 北京科技大学 | Image splicing method and splicing device |
| CN111145161A (en) * | 2019-12-28 | 2020-05-12 | 北京工业大学 | Method for processing and identifying pavement crack digital image |
| CN111310558A (en) * | 2019-12-28 | 2020-06-19 | 北京工业大学 | Pavement disease intelligent extraction method based on deep learning and image processing method |
| CN111551119A (en) * | 2020-05-25 | 2020-08-18 | 广州中恒工程技术有限公司 | Track instrument for detecting irregular crack length of water transport engineering entity |
| CN111665254A (en) * | 2020-06-15 | 2020-09-15 | 陈鹏 | Bridge crack detection method |
| CN111862051A (en) * | 2020-02-04 | 2020-10-30 | 牧今科技 | Method and system for performing automatic camera calibration |
| CN112053331A (en) * | 2020-08-28 | 2020-12-08 | 西安电子科技大学 | Bridge crack detection method based on image superposition and crack information fusion |
| CN112150541A (en) * | 2020-09-10 | 2020-12-29 | 中国石油大学(华东) | Multi-LED wafer positioning algorithm |
| CN112200824A (en) * | 2020-09-30 | 2021-01-08 | 山东省交通科学研究院 | Method for accurately calculating actual width of single pixel in crack image |
| CN112329796A (en) * | 2020-11-12 | 2021-02-05 | 北京环境特性研究所 | Infrared imaging cirrus cloud detection method and device based on visual saliency |
| CN112734720A (en) * | 2021-01-08 | 2021-04-30 | 沈阳工业大学 | Ship hull laser cleaning in-place detection method and system based on visual identification |
| CN112785594A (en) * | 2021-03-16 | 2021-05-11 | 合肥工业大学 | Automatic bridge structure crack identification method based on image two-dimensional amplitude estimation |
| CN112801972A (en) * | 2021-01-25 | 2021-05-14 | 武汉理工大学 | Bridge defect detection method, device, system and storage medium |
| CN113139302A (en) * | 2021-05-20 | 2021-07-20 | 电子科技大学 | Area growth-based solution breaking identification method |
| CN113160202A (en) * | 2021-04-30 | 2021-07-23 | 汕头大学 | Crack detection method and system |
| CN113343976A (en) * | 2021-05-13 | 2021-09-03 | 武汉大学 | Anti-highlight interference engineering measurement mark extraction method based on color-edge fusion feature growth |
| CN113610060A (en) * | 2021-09-29 | 2021-11-05 | 北京雷图科技有限公司 | Structure crack sub-pixel detection method |
| CN113960068A (en) * | 2021-11-23 | 2022-01-21 | 北京华能新锐控制技术有限公司 | Wind power blade damage detection method |
| CN114119614A (en) * | 2022-01-27 | 2022-03-01 | 天津风霖物联网科技有限公司 | Method for remotely detecting cracks of building |
| CN114743106A (en) * | 2022-04-20 | 2022-07-12 | 中钢集团马鞍山矿山研究总院股份有限公司 | Image batch processing identification method and system |
| CN114897892A (en) * | 2022-07-12 | 2022-08-12 | 聊城大学 | Method for calculating characteristic parameters of apparent cracks and holes of PC (polycarbonate) member |
| CN114897899A (en) * | 2022-07-13 | 2022-08-12 | 南通海王消防水带有限公司 | Fire hose low-temperature resistance detection method based on pattern recognition |
| CN115115627A (en) * | 2022-08-29 | 2022-09-27 | 山东省科霖检测有限公司 | Soil saline-alkali soil monitoring method based on data processing |
| US11508088B2 (en) | 2020-02-04 | 2022-11-22 | Mujin, Inc. | Method and system for performing automatic camera calibration |
| CN116030064A (en) * | 2023-03-30 | 2023-04-28 | 南昌工程学院 | Physical crack detection method, system and computer equipment |
| CN116309791A (en) * | 2023-05-17 | 2023-06-23 | 南京星罗基因科技有限公司 | Method for detecting feather area parameters of poultry |
| CN116485801A (en) * | 2023-06-26 | 2023-07-25 | 山东兰通机电有限公司 | Rubber tube quality online detection method and system based on computer vision |
| CN117058129A (en) * | 2023-10-09 | 2023-11-14 | 安徽建筑大学 | Automatic bridge apparent disease identification method based on image processing |
| CN117132602A (en) * | 2023-10-27 | 2023-11-28 | 湖南三昌泵业有限公司 | Visual inspection method for defects of centrifugal pump impeller |
| CN117173661A (en) * | 2023-11-02 | 2023-12-05 | 中铁五局集团成都工程有限责任公司 | Asphalt road quality detection method based on computer vision |
| CN117237330A (en) * | 2023-10-19 | 2023-12-15 | 山东鑫润机电安装工程有限公司 | Automatic bridge defect detection method based on machine vision |
| CN117825408A (en) * | 2024-03-05 | 2024-04-05 | 北京中科蓝图科技有限公司 | Integrated detection method, device and equipment for road |
| WO2024080436A1 (en) * | 2022-10-11 | 2024-04-18 | 주식회사 에프디 | Ai crack-detecting method using bridge pier driving device |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105825169A (en) * | 2016-03-10 | 2016-08-03 | 辽宁工程技术大学 | Road-image-based pavement crack identification method |
| US20170323279A1 (en) * | 2015-05-12 | 2017-11-09 | A La Carte Media Inc. | Kiosks for remote collection of electronic devices for value, and associated mobile application for enhanced diagnostics and services |
| CN107506787A (en) * | 2017-07-27 | 2017-12-22 | 陕西师范大学 | A kind of glue into concrete beam cracks sorting technique based on migration self study |
| CN108038883A (en) * | 2017-12-06 | 2018-05-15 | 陕西土豆数据科技有限公司 | A kind of Crack Detection and recognition methods applied to highway pavement video image |
| CN108611993A (en) * | 2018-05-11 | 2018-10-02 | 杭州电子科技大学 | A kind of modularization Cracks on Concrete Bridge glue-injection machine |
| CN109521019A (en) * | 2018-11-09 | 2019-03-26 | 华南理工大学 | A kind of bridge bottom crack detection method based on unmanned plane vision |
| CN109580657A (en) * | 2019-01-23 | 2019-04-05 | 郑州工程技术学院 | A kind of crack detection method in bridge quality testing |
| CN109615616A (en) * | 2018-11-27 | 2019-04-12 | 北京联合大学 | A kind of crack identification method and system based on ABC-PCNN |
| EP3475914A1 (en) * | 2016-06-28 | 2019-05-01 | Ecoatm, Inc. | Methods and systems for detecting cracks in illuminated electronic device screens |
| CN109754368A (en) * | 2019-01-23 | 2019-05-14 | 郑州工程技术学院 | A kind of crack joining method in bridge quality testing |
-
2019
- 2019-06-26 CN CN201910560732.9A patent/CN110378879B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170323279A1 (en) * | 2015-05-12 | 2017-11-09 | A La Carte Media Inc. | Kiosks for remote collection of electronic devices for value, and associated mobile application for enhanced diagnostics and services |
| CN105825169A (en) * | 2016-03-10 | 2016-08-03 | 辽宁工程技术大学 | Road-image-based pavement crack identification method |
| EP3475914A1 (en) * | 2016-06-28 | 2019-05-01 | Ecoatm, Inc. | Methods and systems for detecting cracks in illuminated electronic device screens |
| CN107506787A (en) * | 2017-07-27 | 2017-12-22 | 陕西师范大学 | A kind of glue into concrete beam cracks sorting technique based on migration self study |
| CN108038883A (en) * | 2017-12-06 | 2018-05-15 | 陕西土豆数据科技有限公司 | A kind of Crack Detection and recognition methods applied to highway pavement video image |
| CN108611993A (en) * | 2018-05-11 | 2018-10-02 | 杭州电子科技大学 | A kind of modularization Cracks on Concrete Bridge glue-injection machine |
| CN109521019A (en) * | 2018-11-09 | 2019-03-26 | 华南理工大学 | A kind of bridge bottom crack detection method based on unmanned plane vision |
| CN109615616A (en) * | 2018-11-27 | 2019-04-12 | 北京联合大学 | A kind of crack identification method and system based on ABC-PCNN |
| CN109580657A (en) * | 2019-01-23 | 2019-04-05 | 郑州工程技术学院 | A kind of crack detection method in bridge quality testing |
| CN109754368A (en) * | 2019-01-23 | 2019-05-14 | 郑州工程技术学院 | A kind of crack joining method in bridge quality testing |
Non-Patent Citations (4)
| Title |
|---|
| YUHUA CHENG 等: "Research on crack detection applications of improved PCNN algorithm in moi nondestructive test method", 《NEUROCOMPUTING》 * |
| ZHIPING CHEN 等: "Analysis on wheel - rail contact and rail cracks of 50m radio telescope", 《ADVANCED MATERIALS RESEARCH》 * |
| 彭博 等: "路面裂缝图像识别算法研究进展", 《中外公路》 * |
| 钟新谷 等: "基于无人飞机成像的桥梁裂缝宽度识别可行性研究", 《土木工程学报》 * |
Cited By (56)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111080514A (en) * | 2019-11-07 | 2020-04-28 | 北京科技大学 | Image splicing method and splicing device |
| CN110969576A (en) * | 2019-11-13 | 2020-04-07 | 同济大学 | Highway pavement image splicing method based on roadside PTZ camera |
| CN111008956A (en) * | 2019-11-13 | 2020-04-14 | 武汉工程大学 | Beam bottom crack detection method, system, device and medium based on image processing |
| CN110991466A (en) * | 2019-11-21 | 2020-04-10 | 同济大学 | Highway road surface condition detecting system based on novel vision sensing equipment |
| CN113780312B (en) * | 2019-11-21 | 2024-04-12 | 同济大学 | Highway road surface condition detecting system |
| CN113780312A (en) * | 2019-11-21 | 2021-12-10 | 同济大学 | Highway road surface condition detecting system |
| CN110926342B (en) * | 2019-11-27 | 2021-07-23 | 北京工业大学 | Crack width measuring method and device |
| CN110926342A (en) * | 2019-11-27 | 2020-03-27 | 北京工业大学 | Crack width measuring method and device |
| CN111145161A (en) * | 2019-12-28 | 2020-05-12 | 北京工业大学 | Method for processing and identifying pavement crack digital image |
| CN111310558A (en) * | 2019-12-28 | 2020-06-19 | 北京工业大学 | Pavement disease intelligent extraction method based on deep learning and image processing method |
| CN111145161B (en) * | 2019-12-28 | 2023-09-19 | 北京工业大学 | Pavement crack digital image processing and identifying method |
| CN111310558B (en) * | 2019-12-28 | 2023-11-21 | 北京工业大学 | Intelligent pavement disease extraction method based on deep learning and image processing method |
| CN111862051A (en) * | 2020-02-04 | 2020-10-30 | 牧今科技 | Method and system for performing automatic camera calibration |
| US11508088B2 (en) | 2020-02-04 | 2022-11-22 | Mujin, Inc. | Method and system for performing automatic camera calibration |
| CN111862051B (en) * | 2020-02-04 | 2021-06-01 | 牧今科技 | Method and system for performing automatic camera calibration |
| CN111551119A (en) * | 2020-05-25 | 2020-08-18 | 广州中恒工程技术有限公司 | Track instrument for detecting irregular crack length of water transport engineering entity |
| CN111665254A (en) * | 2020-06-15 | 2020-09-15 | 陈鹏 | Bridge crack detection method |
| CN112053331B (en) * | 2020-08-28 | 2023-04-07 | 西安电子科技大学 | Bridge crack detection method based on image superposition and crack information fusion |
| CN112053331A (en) * | 2020-08-28 | 2020-12-08 | 西安电子科技大学 | Bridge crack detection method based on image superposition and crack information fusion |
| CN112150541A (en) * | 2020-09-10 | 2020-12-29 | 中国石油大学(华东) | Multi-LED wafer positioning algorithm |
| CN112200824A (en) * | 2020-09-30 | 2021-01-08 | 山东省交通科学研究院 | Method for accurately calculating actual width of single pixel in crack image |
| CN112329796B (en) * | 2020-11-12 | 2023-05-23 | 北京环境特性研究所 | Infrared imaging cloud detection method and device based on visual saliency |
| CN112329796A (en) * | 2020-11-12 | 2021-02-05 | 北京环境特性研究所 | Infrared imaging cirrus cloud detection method and device based on visual saliency |
| CN112734720B (en) * | 2021-01-08 | 2024-03-05 | 沈阳工业大学 | Ship hull laser cleaning in-situ detection method and system based on visual identification |
| CN112734720A (en) * | 2021-01-08 | 2021-04-30 | 沈阳工业大学 | Ship hull laser cleaning in-place detection method and system based on visual identification |
| CN112801972A (en) * | 2021-01-25 | 2021-05-14 | 武汉理工大学 | Bridge defect detection method, device, system and storage medium |
| CN112785594A (en) * | 2021-03-16 | 2021-05-11 | 合肥工业大学 | Automatic bridge structure crack identification method based on image two-dimensional amplitude estimation |
| CN112785594B (en) * | 2021-03-16 | 2022-08-30 | 合肥工业大学 | Automatic bridge structure crack identification method based on image two-dimensional amplitude estimation |
| CN113160202A (en) * | 2021-04-30 | 2021-07-23 | 汕头大学 | Crack detection method and system |
| CN113343976A (en) * | 2021-05-13 | 2021-09-03 | 武汉大学 | Anti-highlight interference engineering measurement mark extraction method based on color-edge fusion feature growth |
| CN113139302A (en) * | 2021-05-20 | 2021-07-20 | 电子科技大学 | Area growth-based solution breaking identification method |
| CN113610060A (en) * | 2021-09-29 | 2021-11-05 | 北京雷图科技有限公司 | Structure crack sub-pixel detection method |
| CN113960068A (en) * | 2021-11-23 | 2022-01-21 | 北京华能新锐控制技术有限公司 | Wind power blade damage detection method |
| CN114119614A (en) * | 2022-01-27 | 2022-03-01 | 天津风霖物联网科技有限公司 | Method for remotely detecting cracks of building |
| CN114743106A (en) * | 2022-04-20 | 2022-07-12 | 中钢集团马鞍山矿山研究总院股份有限公司 | Image batch processing identification method and system |
| CN114743106B (en) * | 2022-04-20 | 2022-11-25 | 中钢集团马鞍山矿山研究总院股份有限公司 | Image batch processing identification method and system |
| CN114897892A (en) * | 2022-07-12 | 2022-08-12 | 聊城大学 | Method for calculating characteristic parameters of apparent cracks and holes of PC (polycarbonate) member |
| CN114897899B (en) * | 2022-07-13 | 2022-09-30 | 南通海王消防水带有限公司 | Fire hose low-temperature resistance detection method based on pattern recognition |
| CN114897899A (en) * | 2022-07-13 | 2022-08-12 | 南通海王消防水带有限公司 | Fire hose low-temperature resistance detection method based on pattern recognition |
| CN115115627B (en) * | 2022-08-29 | 2022-11-15 | 山东省科霖检测有限公司 | Soil saline-alkali soil monitoring method based on data processing |
| CN115115627A (en) * | 2022-08-29 | 2022-09-27 | 山东省科霖检测有限公司 | Soil saline-alkali soil monitoring method based on data processing |
| WO2024080436A1 (en) * | 2022-10-11 | 2024-04-18 | 주식회사 에프디 | Ai crack-detecting method using bridge pier driving device |
| CN116030064A (en) * | 2023-03-30 | 2023-04-28 | 南昌工程学院 | Physical crack detection method, system and computer equipment |
| CN116309791A (en) * | 2023-05-17 | 2023-06-23 | 南京星罗基因科技有限公司 | Method for detecting feather area parameters of poultry |
| CN116309791B (en) * | 2023-05-17 | 2023-10-27 | 南京星罗基因科技有限公司 | Method for detecting feather area parameters of poultry |
| CN116485801A (en) * | 2023-06-26 | 2023-07-25 | 山东兰通机电有限公司 | Rubber tube quality online detection method and system based on computer vision |
| CN116485801B (en) * | 2023-06-26 | 2023-09-12 | 山东兰通机电有限公司 | Rubber tube quality online detection method and system based on computer vision |
| CN117058129B (en) * | 2023-10-09 | 2024-01-12 | 安徽建筑大学 | Automatic bridge apparent disease identification method based on image processing |
| CN117058129A (en) * | 2023-10-09 | 2023-11-14 | 安徽建筑大学 | Automatic bridge apparent disease identification method based on image processing |
| CN117237330A (en) * | 2023-10-19 | 2023-12-15 | 山东鑫润机电安装工程有限公司 | Automatic bridge defect detection method based on machine vision |
| CN117237330B (en) * | 2023-10-19 | 2024-02-20 | 山东鑫润机电安装工程有限公司 | Automatic bridge defect detection method based on machine vision |
| CN117132602B (en) * | 2023-10-27 | 2024-01-02 | 湖南三昌泵业有限公司 | Visual inspection method for defects of centrifugal pump impeller |
| CN117132602A (en) * | 2023-10-27 | 2023-11-28 | 湖南三昌泵业有限公司 | Visual inspection method for defects of centrifugal pump impeller |
| CN117173661B (en) * | 2023-11-02 | 2024-01-26 | 中铁五局集团成都工程有限责任公司 | Asphalt road quality detection method based on computer vision |
| CN117173661A (en) * | 2023-11-02 | 2023-12-05 | 中铁五局集团成都工程有限责任公司 | Asphalt road quality detection method based on computer vision |
| CN117825408A (en) * | 2024-03-05 | 2024-04-05 | 北京中科蓝图科技有限公司 | Integrated detection method, device and equipment for road |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110378879B (en) | 2021-03-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110378879A (en) | A kind of Bridge Crack detection method | |
| CN110390669A (en) | The detection method in crack in a kind of bridge image | |
| Chen et al. | A self organizing map optimization based image recognition and processing model for bridge crack inspection | |
| CN104792792B (en) | A kind of road surface crack detection method of Stepwise Refinement | |
| CN104021574B (en) | Pavement disease automatic identifying method | |
| CN108596165B (en) | Road traffic marking detection method and system based on unmanned plane low latitude Aerial Images | |
| CN103279765B (en) | Steel wire rope surface damage detection method based on images match | |
| CN105761271A (en) | Nuclear safety shell surface defect automatic detection method and system | |
| US20010036293A1 (en) | System for automatically generating database of objects of interest by analysis of images recorded by moving vehicle | |
| CN110569730B (en) | Road surface crack automatic identification method based on U-net neural network model | |
| CN101782526B (en) | Method and device for automatically restoring, measuring and classifying steel dimple images | |
| CN105389586A (en) | Method for automatically detecting integrity of shrimp body based on computer vision | |
| CN106250936A (en) | Multiple features multithreading safety check contraband automatic identifying method based on machine learning | |
| CN101877074A (en) | Tubercle bacillus target recognizing and counting algorithm based on diverse characteristics | |
| CN108416774A (en) | A kind of fabric types recognition methods based on fine granularity neural network | |
| CN108921076A (en) | Road face crack disease self-adaption constant false-alarm detection method based on image | |
| CN112964712A (en) | Method for rapidly detecting state of asphalt pavement | |
| CN102855485B (en) | The automatic testing method of one grow wheat heading | |
| CN107729853A (en) | A kind of automatic identifying method suitable for the narrow tuning drive gear formula instrument of transformer station | |
| CN104992429A (en) | Mountain crack detection method based on image local reinforcement | |
| CN106023226A (en) | Crack automatic detection method based on three-dimensional virtual pavement | |
| CN115690081A (en) | Tree counting method, system, storage medium, computer equipment and terminal | |
| CN110503637A (en) | A kind of crack on road automatic testing method based on convolutional neural networks | |
| CN113313107A (en) | Intelligent detection and identification method for multiple types of diseases on cable surface of cable-stayed bridge | |
| CN110363706A (en) | A kind of large area bridge floor image split-joint method |
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 |