CN109187548A - A kind of rock cranny recognition methods - Google Patents

A kind of rock cranny recognition methods Download PDF

Info

Publication number
CN109187548A
CN109187548A CN201810817958.8A CN201810817958A CN109187548A CN 109187548 A CN109187548 A CN 109187548A CN 201810817958 A CN201810817958 A CN 201810817958A CN 109187548 A CN109187548 A CN 109187548A
Authority
CN
China
Prior art keywords
line
section
mod
subset
value
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
Application number
CN201810817958.8A
Other languages
Chinese (zh)
Other versions
CN109187548B (en
Inventor
冷彪
崔兆东
杨辉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southwest Jiaotong University
Original Assignee
Southwest Jiaotong University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Southwest Jiaotong University filed Critical Southwest Jiaotong University
Priority to CN201810817958.8A priority Critical patent/CN109187548B/en
Publication of CN109187548A publication Critical patent/CN109187548A/en
Application granted granted Critical
Publication of CN109187548B publication Critical patent/CN109187548B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
    • 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
    • G06T7/13—Edge detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/181—Segmentation; Edge detection involving edge growing; involving edge linking
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
    • G01N2021/8854—Grading and classifying of flaws
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
    • G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
    • 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
    • G06T2207/10—Image acquisition modality
    • G06T2207/10004—Still image; Photographic image
    • 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
    • G06T2207/10—Image acquisition modality
    • G06T2207/10024—Color image
    • 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)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Signal Processing (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Quality & Reliability (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

The invention discloses a kind of rock cranny recognition methods, including obtain multi-section-line, multi-section-line separation to rock mass face image progress edge detection, image thinning and removing node of divergence, extraction boundary line, fitting, according to length threshold removing multi-section-line, the similar multi-section-line of merging and again according to length threshold removing multi-section-line.Further technical solution further includes counting all multi-section-line angles and being grouped according to the distribution situation of angle value to crack.The beneficial effects of the present invention are non local progress crack identification and extractions to complete tunnel tunnel face；Discontinuous crack is connected, is allowed to more complete；The crack extracted is grouped, the rock cranny boundary line face geologic sketch map under different grouping is obtained.

Description

A kind of rock cranny recognition methods
Technical field
The invention belongs to rock cranny application field, tunnel tunnel face is completed in especially a kind of rock cranny recognition methods Upper rock cranny extracts and grouping.
Background technique
The information content that tunnel tunnel face contains is huge, and face rock mass conditions are evaluation Analyses of Tunnel Wall Rock Stability, determine tunnel One of road supporting scheme and the important evidence of construction technology.It in Practical Project, is limited by field condition, technical conditions, many feelings Still through technical staff under condition, according to set record format, the faithful record fills in face geological information.Such as fill in personnel without Practical experience often results in and fills in leakage item or make a fault, and leads to being out of one's reckoning or making a fault for fender graded.
For this case, some scholars automatically extract face geological information, analysis method is studied.At present There are mainly two types of the data sources of automated information retrieval: (1) digital photographing (2) 3 D laser scanning.Digital photographing is obtained There are mainly two types of processing methods for digital picture, one is directly Digital Image Processing analysis is carried out to digital photograph, such as: using Image processing algorithm extracts rock cranny, automatically parses acquisition wall-rock crack development degree parameter according to result is extracted；Directly Statistics characteristic analysis is carried out to pretreated image.Another kind is to be clapped based on close-shot photography measure technique according to different angle The mapping relations of datum mark in the face photo taken the photograph calculate three-dimensional space point cloud data, then analyze rock cranny length, rock The information such as layer occurrence.The three-dimensional laser point cloud data obtained using 3 D laser scanning then may be directly applied to analysis rock mass knot Structure region feature information.
At present apparently, three-dimensional laser point cloud has spatial character, the palm based on close-range photogrammetry and 3 D laser scanning Sub- surface treatment can preferably realize the extraction to structural surface information, but its degree of automation is still inadequate, it is also necessary to manual intervention It just can apply to reality.Area surface treatment based on digital image processing techniques is also very immature, very due to face image Complexity, correlative study are mostly the processing to some ideal images or face local area image, it is difficult to be applied to practical work Journey.But the acquisition of face image has the characteristics that relatively easy, the time is short, equipment manufacturing cost is low, the processing point to face image The related algorithm research of analysis is still very necessary.
Summary of the invention
The purpose of the present invention is to propose to a kind of rock cranny recognition methods, realize extraction to rock cranny on face and Grouping.
The technical solution achieved the object of the present invention is as follows:
A kind of rock cranny recognition methods, including
Step 1: edge detection is carried out to rock mass face image；
Step 2: boundary line processing, including image thinning and removing node of divergence are carried out to rock mass face image；
Step 3: boundary line is extracted；
Step 4: being fitted to obtain multi-section-line to each boundary line extracted, including
4.1 pixels for enabling a line boundary line pass through are respectively pi, i=1,2 ..., n, wherein p1And pnFor The endpoint of multi-section-line；
4.2 by p1As the starting point of multi-section-line, and current pixel point is set as pj, successively calculate p1To pjBetween each pixel Point arrives line segment p1pjDistance；If p1To pjBetween, there are more than one pixels to line segment p1pjDistance value be greater than or equal to Given value dT, then by pj-1As an endpoint on multi-section-line, then with pj-1For starting point, continued point using same method The pixel of analysis thereafter, until finding out all endpoints on multi-section-line；Otherwise, current pixel point is revised as pj+1, by aforementioned Method continues to search endpoint；
4.3 are sequentially connected each endpoint, obtain multi-section-line；
Step 5: each multi-section-line is separated according to corner dimension between its adjacent segments: setting separation threshold alphaT, αTLess than 180 degree；If angle is less than α between two adjacent segments on multi-section-lineT, then by the multi-section-line from this two adjacent segments Two multi-section-lines are separated at public point；
Step 6: setting length threshold lT, remove in all multi-section-lines, the sum of each line segment length is less than length threshold lT's Multi-section-line；
Step 7: merging similar multi-section-line, including
7.1 choose a multi-section-line as current multi-section-line, and successively each multi-section-line adjacent with its periphery is compared, As follows: the arest neighbors endpoint wire length D of condition 1, current multi-section-line and adjacent multi-section-line is less than given threshold value DT；Condition 2, The angle β of arest neighbors endpoint line and current multi-section-line1Not less than given angle threshold value betaT；Condition 3: arest neighbors endpoint line with The angle β of adjacent multi-section-line2Not less than given angle threshold value betaT；As met above-mentioned 3 conditions simultaneously, then the adjacent multi-section-line is enabled Similarity value R=min (β1,β2)/D；
7.2 take the maximum value R of similarity value RmaxCorresponding adjacent multi-section-line and current multi-section-line are from arest neighbors endpoint It is connected, is merged into a multi-section-line；
7.3 choose remaining multi-section-line successively as current multi-section-line, are handled according to the method for 7.1 and 7.2；
Step 8: removing in all multi-section-lines, and the sum of each line segment length is less than length threshold LTMulti-section-line.
It further, further include counting all multi-section-line angles and being divided according to the distribution situation of angle value crack Group:
If i-th multi-section-line is L on facei, length li, have n line segment thereon, each line segment length is li,j, j= 1,2 ..., n, the ray R of each line segment and direction horizontally to the righti,jBetween angle be γi,jIf final multi-section-line grouping set is G；
(1) multi-section-line L is calculatediWith the ray R in direction horizontally to the rightiBetween angle, if multi-section-line LiIt is penetrated with direction horizontally to the right Line RiAngle is γi, then
(2) according to multi-section-line γi[0 °, 10 °), [10 °, 20 °) ..., [170 °, 180 °) in corresponding angular range, All multi-section-lines are grouped, grouping set { P is obtainedk| k=0,1 ..., 17 }, PkIndicate the multistage in same angular range Line set；
(3) set of computations { Pk| k=0,1 ..., 17 } each subset PkThe sum of multistage line length sk, gathered { sk| k= 0,1,…,17}；Enable k1=(k-1+18) mod 18, k2=(k+1) mod 18, if sk≥sk1And sk≥sk2, then the point is recorded K value and skValue；By all k values for meeting condition and its corresponding skIt is worth the point pair constituted, sorts by k value, be saved in set {(kl,sl) | l=0,1 ..., L-1, lm<lm+1In；
(4) l1=(l+1) mod L is enabled, if | kl-kl1|≤3, then by (kl,sl) and (kl1,sl1) sequentially it is merged into same collection In conjunction；{ (k will be gatheredl,sl) | l=0,1 ..., L-1, kl<kl+1In all elements for meeting condition, be respectively merged into correspondence In set, { T finally is added as subset in all setm| m=0,1 ..., M-1 }, Tm={ (km,n,sm,n) | n=0,1 ..., N-1}；
(5) { T is successively analyzedm| m=0,1 ..., M-1 in subset Tm；If current TmHead and the tail element k value be respectively km,0 And km,N-1, then
1. successively enabling K1=(km,0-o1+18)mod 18(o1=1,2 ..., 17), enable K2=(K1-1+18) mod 18, K3 =(K1+1) mod 18, if in set { sk| k=0,1 ..., 17 } in, sK1< sK2And sK1< sK3, then enableInto Step is 2.；
2. successively enabling K4=(km,N-1+o2)mod 18(o2=1,2 ..., 17), enable K5=(K4-1+18) mod 18, K6= (K4+1) 18 mod, if in set { sk| k=0,1 ..., 17 } in, sK4< sK5And sK4< sK6, then enableInto step Suddenly 3.；
3. if InElement number is enabled less than 3In set { Pk| k=0,1 ..., 17 } In, it finds out respectivelyCorresponding subset PkIn multi-section-line, all merge into one Subset is saved in set G, and return step (5) continues to execute；
4. if InElement number is not less than 3, in InIn successively analyze 3 continuous element (kn,a1-1,sn,a1-1)、(kn,a1, sn,a1) and (kn,a1+1, sn,a1+1),
If a, meeting sn,a1<sn,a1-1And sn,a1<sn,a1+1, in set { (kl,sl) | l=0,1 ..., L-1 } in, if with (kn,a1-1,sn,a1-1)、(kn,a1+1, sn,a1+1) corresponding element is respectively (km1,sm1)、(km2,sm2)；Enable Δ k1=(km2-km1+ 18) 18 mod, in set { sk| k=0,1 ..., 17 } in, search k=(km1+o4)mod 18(o4=1,2 ..., Δ k1-1) when Minimum in corresponding all elements, if the element numbers areSimilarly, I is found outnIn all minimums
In set { Pk| k=0,1 ..., 17 } in, it finds outIt arrivesBetween multi-section-line set, it may be assumed that enable It finds out respectivelyCorresponding subset PkIn Multi-section-line is all merged into a subset and is saved in set G；Similarly, it finds out respectivelyIt arrivesIt arrivesBetween multi-section-line formed subset, be saved in set G；Return step (5) continues to execute；
If b, meeting the condition of a without 3 continuous elements, in set { Pk| k=0,1 ..., 17 } in, it finds out serial number and existsIt arrivesBetween subset multi-section-line, formed a subset, be saved in set G；Return step (5) continues to execute；
Step (5) has analyzed { Im| m=0,1 ..., M-1 in all subset ImAfterwards, the set G obtained is ultimate bound Line group result.
The beneficial effects of the present invention are,
One, non local progress crack identification and extraction to complete tunnel tunnel face.
Two, discontinuous crack is connected, is allowed to more complete.
Three, the crack extracted is grouped, with obtaining the rock cranny boundary line face under different grouping quality Tracing.
Detailed description of the invention
Fig. 1 is that rock cranny extracts and be grouped flow chart on tunnel tunnel face in the present invention；
Fig. 2 is original boundaries and polymerization schematic diagram in the present invention；
Fig. 3 is fitting multi-section-line seperated schematic diagram in the present invention；
Fig. 4 is two to close on multi-section-line similarity analysis schematic diagram in the present invention；
Fig. 5 is face image and rock cranny Boundary Extraction and group result figure.
Specific embodiment
The process flow of face image is as shown in Figure 1.Firstly, allocating conventional algorithm pre-processes face image, as possible It is improved picture quality；Then edge detection and boundary line drawing are carried out to image, and boundary is carried out on this basis Fitting splits and connects, and obtains rock cranny boundary line and automatically extracts result；After extracting result progress artificial correction, use Algorithm is grouped crack boundary line automatically, if group result is wrong, can carry out manually being grouped amendment again, finally obtain Rock cranny boundary line face geologic sketch map under different grouping.
1, image preprocessing
The main purpose of image preprocessing is to increase rock cranny edge and ambient background color contrast as possible, is reduced non- Gap region color contrast is more easier the subsequent identification to rock cranny.
Tunnel tunnel face image is affected by many factors in collection process, will unavoidably picture quality be caused to reduce. Dust of such as constructing influences, illumination is insufficient, uneven illumination is even, the time for exposure is too long.Imaging of the different factors to face photo Quality influences also difference, and this requires the preprocess method to different images is different.The image preprocessing being currently known is calculated There are many method and related software, completely can according to operator's use experience, by adjusting brightness/contrast, sharpening/softening, Exposure etc. keeps brightness of image moderate, and rock cranny and periphery color contrast are obvious, convenient for passing through edge detection algorithm Realize the detection to rock cranny.
2, edge detection
Edge detection can will test out the rock cranny on face, to extract and identifying that base is laid on rock cranny boundary Plinth.
2.1 edge detection
Traditional edge detection operator has Roberts, Prewitt, Sobel, Log operator etc., although can play certain Edge detection effect, but the main feature of these operators is to noise-sensitive, and edge precision is low.And Canny operator can use up Amount reduces influence of noise, identifies actual edge as much as possible, and can make the edge detected and actual edge maximum journey Degree approaches.Canny operator testing result greatly remains true rock cranny edge, and details is richer, and connectivity is good, This will greatly improve success rate to rock cranny Boundary Extraction, and the rock cranny edge continuity that other operators detect Difference, it is difficult to form more complete Edge Feature Points.Therefore selection Canny operator carries out edge detection to face image.
The processing of 2.2 boundary lines
The purpose of Boundary Extraction is that all edge extractings comprising rock cranny for obtaining edge detection come out, to analyze With the real rock cranny edge of identification.
2.2.1 image thinning
Image thinning can make all edges comprising rock cranny become single pixel wide boundary, in order to extract boundary Line.For the edge that Canny operator detects, extra edge pixel point can be removed by image thinning.
When refining edge, it can judge whether the point can remove according to 8 neighborhood situations of a point.The refinement principle of use Are as follows:
(1) if the internal point of cut zone after deleting, which cannot be deleted；
(2) if can shorten edge after deleting, which is straight line endpoint, cannot be deleted；
(3) if can destroy connectivity after deleting, which cannot be deleted.
2.2.2 node of divergence is removed
There are bifurcated, possible reasons for boundary line target after face image thinning are as follows:
(1) crack is mutually completed a business transaction
(2) rock crushing
(3) uneven color
(4) mechanical execution trace
(5) picture noise
By removing the node of divergence on boundary line, the largely boundary line target without node of divergence is formed, boundary line is analyzed Target signature can be conducive to extract and identify real rock cranny.
3, boundary line secondary treatment
Side is extracted in the face rock mass image for refining and eliminating crossover node and is analyzed in boundary line secondary treatment Boundary line automatically identifies possible rock cranny.
3.1 Boundary Extraction
In all face rock mass images for being refined and being eliminated crossover node, same boundary line target pixel points exist It is continuously, all boundary line targets to be gone out according to this feature extraction in 8 connected regions.
3.2 edge fitting
In order to find out possible rock cranny, boundary line is subjected to multi-section-line fitting.
If the pixel that a line boundary line passes through is respectively pi(i=1,2 ..., n), by p1It is more as being fitted Section line starting point, each pixel is successively analyzed along boundary line.If current pixel point is pj, successively calculate p1To pjBetween each pixel To line segment p1pjDistance.If all pixels point is to line segment p1pjDistance value be respectively less than given value dT, then continue to analyze pj+1；Instead It, then by pj-1It is saved as the endpoint on fitting multi-section-line, then with pj-1For starting point, continue to analyze using same method Thereafter pixel, until finding the endpoint on all fitting multi-section-lines.
As shown in Fig. 2, pixel p1、p7Between pixel to line segment p1p7Distance be less than given value dT, but pixel Point p1、p8Between certain pixel to line segment p1p8Distance be more than dT, therefore by p7As an endpoint on fitting multi-section-line, together Reason obtains p12、p16, with p1、p7、p12、p16Multi-section-line for endpoint is fitting result.
3.3 boundaries are split
All boundary line targets extracted, it may be possible to smoother straight line, it is also possible to very curved curve, and rock Body crack is usually near linear in smaller range.Therefore boundary line target has 3 kinds of possibility:
(1) rock cranny is fully belonged to
(2) rock cranny is partly belonged to
(3) it is not belonging to rock cranny
After boundary line is fitted to multi-section-line, the fitting multi-section-line of situation (1) is near linear；The fitting multistage of situation (2) Line is belonging to rock cranny part, in approximate relatively long straight line；The fitting multi-section-line of situation (3) is without longer near linear.
It is more by being fitted in order to distinguish rock cranny boundary that may be present and non-rock cranny boundary on same multi-section-line Section line carries out separation according to corner dimension situation between its adjacent segments and is filtered with to the boundary line after sharing:
(1) setting separation threshold alphaTIf angle is less than α between certain two adjacent segments on multi-section-lineT, then by multi-section-line from this two Two multi-section-lines are separated at the common point of adjacent segments.Multi-section-line p in Fig. 31p2p3p4p5Two adjacent two line segments p2p3With p3p4Angle α be less than αT(αT< 180 °), then in p3Multi-section-line is separated into p by place1p2p3And p3p4p5Two multi-section-lines.
(2) after all boundary lines are split, length threshold l is setTEach multi-section-line is screened, all length is removed and is less than lTMulti-section-line.
3.4 contour connection
In face image after all boundary line fitting multi-section-line separation, according to rock cranny in close in smaller area Like the characteristics of straight line, compared with each multi-section-line is carried out two-by-two with adjacent multi-section-line, two multi-section-lines are analyzed in same rock mass A possibility that on crack, here referred to as the similitude of adjacent multi-section-line.
Degree of similarity is accounted in terms of two: (1) distance of two multi-section-line arest neighbors two-end-points and (2) this both ends The angle of point line and two line segment of arest neighbors.
If the arest neighbors endpoint wire length D of adjacent multi-section-line (see Fig. 4) is less than given threshold value D two-by-twoT, line with it is adjacent The angle β of two line segments1、β2It is not less than given angle threshold value betaT(βT< 180 °), it enables
R=min (β1,β2)/D(β1,β2≥βT, and D < DT)
R value reflects the similarity degree of two adjacent multi-section-lines.By all multi-section-lines that may be connected with current multi-section-line side R value be compared, R value is maximized RmaxCorresponding multi-section-line is connected from minimum distance endpoint with current multi-section-line, It is merged into a multi-section-line.The multi-section-line is possible rock cranny edge fitting multi-section-line.
After multi-section-line connection, multi-section-line is screened again, length threshold L is setT, remove all length and be less than LT's Multi-section-line.
4, boundary line is grouped automatically
In same group of rock cranny, it is almost the same to be fitted multi-section-line direction.All multi-section-line angles are counted, it can basis The distribution situation of angle value is grouped crack.
If i-th multi-section-line is L on facei, length li, have n line segment thereon, each line segment length is li,j(j= 1,2 ..., n), the ray R of each line segment and direction horizontally to the righti,jBetween angle be γi,jIf final multi-section-line grouping set is G.
(1) multi-section-line L is calculatediWith the ray R in direction horizontally to the rightiBetween angle, if multi-section-line LiIt is penetrated with direction horizontally to the right Line RiAngle is γi, then
(2) according to multi-section-line γi[0 °, 10 °), [10 °, 20 °) ..., [170 °, 180 °) in corresponding angular range, All multi-section-lines are grouped, grouping set { P is obtainedk| k=0,1 ..., 17 }, PkIndicate the multistage in same angular range Line set；
(3) set of computations { Pk| k=0,1 ..., 17 } each subset PkThe sum of multistage line length sk, gathered { sk| k= 0,1,…,17}.Enable k1=(k-1+18) mod 18, k2=(k+1) mod 18, if sk≥sk1And sk≥sk2, then the point is recorded K value and skValue.By all k values for meeting condition and its corresponding skIt is worth the point pair constituted, sorts by k value, be saved in set {(kl,sl) | l=0,1 ..., L-1, lm<lm+1In.
(4) l1=(l+1) mod L is enabled, if | kl-kl1|≤3, then by (kl,sl) and (kl1,sl1) sequentially it is merged into same collection In conjunction.{ (k will be gatheredl,sl) | l=0,1 ..., L-1, kl<kl+1In all elements for meeting condition, be respectively merged into correspondence In set, { T finally is added as subset in all setm| m=0,1 ..., M-1 }, Tm={ (km,n,sm,n) | n=0,1 ..., N-1}。
(5) { T is successively analyzedm| m=0,1 ..., M-1 in subset Tm.If current TmHead and the tail element k value be respectively km,0 And km,N-1, then
1. successively enabling K1=(km,0-o1+18)mod 18(o1=1,2 ..., 17), enable K2=(K1-1+18) mod 18, K3 =(K1+1) mod 18, if in set { sk| k=0,1 ..., 17 } in, sK1< sK2And sK1< sK3, then enableInto Step is 2.；
2. successively enabling K4=(km,N-1+o2)mod 18(o2=1,2 ..., 17), enable K5=(K4-1+18) mod 18, K6= (K4+1) 18 mod, if in set { sk| k=0,1 ..., 17 } in, sK4< sK5And sK4< sK6, then enableInto step Suddenly 3.；
3. if InElement number is enabled less than 3In set { Pk| k=0,1 ..., 17 } In, it finds out respectivelyCorresponding subset PkIn multi-section-line, all merge into one A subset is saved in set G, and return step (5) continues to execute；
4. if InElement number is not less than 3, in InIn successively analyze 3 continuous element (kn,a1-1,sn,a1-1)、(kn,a1, sn,a1) and (kn,a1+1, sn,a1+1),
If a, meeting sn,a1<sn,a1-1And sn,a1<sn,a1+1, in set { (kl,sl) | l=0,1 ..., L-1 } in, if with (kn,a1-1,sn,a1-1)、 (kn,a1+1,sn,a1+1) corresponding element is respectively (km1,sm1)、(km2,sm2).Enable Δ k1=(km2-km1+ 18) 18 mod, in set { sk| k=0,1 ..., 17 } in, search k=(km1+o4)mod 18(o4=1,2 ..., Δ k1-1) when Minimum in corresponding all elements, if the element numbers areSimilarly, I is found outnIn all minimums
In set { Pk| k=0,1 ..., 17 } in, it finds outIt arrivesBetween multi-section-line set, it may be assumed that enable It finds out respectivelyCorresponding subset PkIn Multi-section-line is all merged into a subset and is saved in set G；Similarly, it finds out respectivelyIt arrivesIt arrivesBetween multi-section-line formed subset, be saved in set G.Return step (5) continues to execute；
If b, meeting the condition of a without 3 continuous elements, in set { Pk| k=0,1 ..., 17 } in, it finds out serial number and existsIt arrivesBetween subset multi-section-line, formed a subset, be saved in set G.Return step (5) continues to execute.
Step (5) has analyzed { Im| m=0,1 ..., M-1 in all subset ImAfterwards, the set G obtained is ultimate bound The automatic group result of line.

Claims (2)

1. a kind of rock cranny recognition methods, which is characterized in that including
Step 1: edge detection is carried out to rock mass face image；
Step 2: boundary line processing, including image thinning and removing node of divergence are carried out to rock mass face image；
Step 3: boundary line is extracted；
Step 4: being fitted to obtain multi-section-line to each boundary line extracted, including
4.1 pixels for enabling a line boundary line pass through are respectively pi, i=1,2 ..., n, wherein p1And pnFor multistage The endpoint of line；
4.2 by p1As the starting point of multi-section-line, and current pixel point is set as pj, successively calculate p1To pjBetween each pixel to line Section p1pjDistance；If p1To pjBetween, there are more than one pixels to line segment p1pjDistance value be greater than or equal to given value dT, then by pj-1As an endpoint on multi-section-line, then with pj-1For starting point, continue to analyze thereafter using same method Pixel, until finding out all endpoints on multi-section-line；Otherwise, current pixel point is revised as pj+1, continue to look by preceding method Look for endpoint；
4.3 are sequentially connected each endpoint, obtain multi-section-line；
Step 5: each multi-section-line is separated according to corner dimension between its adjacent segments: setting separation threshold alphaT, αTIt is small In 180 degree；If angle is less than α between two adjacent segments on multi-section-lineT, then by the multi-section-line from the public of this two adjacent segments Two multi-section-lines are separated at endpoint；
Step 6: setting length threshold lT, remove in all multi-section-lines, the sum of each line segment length is less than length threshold lTMultistage Line；
Step 7: merging similar multi-section-line, including
7.1 choose a multi-section-line as current multi-section-line, and successively each multi-section-line adjacent with its periphery is compared, such as Under: the arest neighbors endpoint wire length D of condition 1, current multi-section-line and adjacent multi-section-line is less than given threshold value DT；Condition 2, recently The angle β of neighboring terminal point line and current multi-section-line1Not less than given angle threshold value betaT；Condition 3: arest neighbors endpoint line with it is adjacent The angle β of multi-section-line2Not less than given angle threshold value betaT；As met above-mentioned 3 conditions simultaneously, then the phase of the adjacent multi-section-line is enabled Like angle value R=min (β1,β2)/D；
7.2 take the maximum value R of similarity value RmaxCorresponding adjacent multi-section-line and current multi-section-line are from phase from arest neighbors endpoint Even, it is merged into a multi-section-line；
7.3 choose remaining multi-section-line successively as current multi-section-line, are handled according to the method for 7.1 and 7.2；
Step 8: removing in all multi-section-lines, and the sum of each line segment length is less than length threshold LTMulti-section-line.
2. a kind of rock cranny recognition methods as described in claim 1, which is characterized in that further include counting all multi-section-lines Angle is simultaneously grouped crack according to the distribution situation of angle value:
If i-th multi-section-line is L on facei, length li, have n line segment thereon, each line segment length is li,j, j=1, 2 ..., n, the ray R of each line segment and direction horizontally to the righti,jBetween angle be γi,jIf final multi-section-line grouping set is G；
(1) multi-section-line L is calculatediWith the ray R in direction horizontally to the rightiBetween angle, if multi-section-line LiWith oriented radial R horizontally to the righti Angle is γi, then
(2) according to multi-section-line γi[0 °, 10 °), [10 °, 20 °) ..., [170 °, 180 °) in corresponding angular range, to all Multi-section-line is grouped, and obtains grouping set { Pk| k=0,1 ..., 17 }, PkIndicate the multi-section-line set in same angular range；
(3) set of computations { Pk| k=0,1 ..., 17 } each subset PkThe sum of multistage line length sk, gathered { sk| k=0, 1,…,17}；Enable k1=(k-1+18) mod 18, k2=(k+1) mod 18, if sk≥sk1And sk≥sk2, then the k of the point is recorded Value and skValue；By all k values for meeting condition and its corresponding skIt is worth the point pair constituted, sorts by k value, be saved in set { (kl, sl) | l=0,1 ..., L-1, lm<lm+1In；
(4) l1=(l+1) mod L is enabled, if | kl-kl1|≤3, then by (kl,sl) and (kl1,sl1) sequentially it is merged into identity set In；{ (k will be gatheredl,sl) | l=0,1 ..., L-1, kl<kl+1In all elements for meeting condition, be respectively merged into corresponding set In, { T finally is added as subset in all setm| m=0,1 ..., M-1 }, Tm={ (km,n,sm,n) | n=0,1 ..., N-1 }；
(5) { T is successively analyzedm| m=0,1 ..., M-1 in subset Tm；If current TmHead and the tail element k value be respectively km,0With km,N-1, then
1. successively enabling K1=(km,0-o1+18)mod 18(o1=1,2 ..., 17), enable K2=(K1-1+18) mod 18, K3=(K1 + 1) 18 mod, if in set { sk| k=0,1 ..., 17 } in, sK1< sK2And sK1< sK3, then enableIt enters step 2.；
2. successively enabling K4=(km,N-1+o2)mod 18(o2=1,2 ..., 17), enable K5=(K4-1+18) mod 18, K6=(K4+ 1) 18 mod, if in set { sk| k=0,1 ..., 17 } in, sK4< sK5And sK4< sK6, then enableIt enters step 3.；
3. if InElement number is enabled less than 3In set { Pk| k=0,1 ..., 17 } in, point It does not find outCorresponding subset PkIn multi-section-line, all merge into a subset guarantor It is stored in set G, return step (5) continues to execute；
4. if InElement number is not less than 3, in InIn successively analyze 3 continuous element (kn,a1-1,sn,a1-1)、(kn,a1,sn,a1) and (kn,a1+1,sn,a1+1),
If a, meeting sn,a1<sn,a1-1And sn,a1<sn,a1+1, in set { (kl,sl) | l=0,1 ..., L-1 } in, if with (kn,a1-1, sn,a1-1)、(kn,a1+1,sn,a1+1) corresponding element is respectively (km1,sm1)、(km2,sm2)；Enable Δ k1=(km2-km1+18)mod 18, in set { sk| k=0,1 ..., 17 } in, search k=(km1+o4)mod 18(o4=1,2 ..., Δ k1-1) when corresponding institute There is the minimum in element, if the element numbers areSimilarly, I is found outnIn all minimums
In set { Pk| k=0,1 ..., 17 } in, it finds outIt arrivesBetween multi-section-line set, it may be assumed that enable It finds out respectivelyCorresponding subset PkIn multi-section-line, all close And it is saved in set G for a subset；Similarly, it finds out respectivelyIt arrivesIt arrivesBetween multi-section-line The subset of formation is saved in set G；Return step (5) continues to execute；
If b, meeting the condition of a without 3 continuous elements, in set { Pk| k=0,1 ..., 17 } in, it finds out serial number and exists It arrivesBetween subset multi-section-line, formed a subset, be saved in set G；Return step (5) continues to execute；
Step (5) has analyzed { Im| m=0,1 ..., M-1 in all subset ImAfterwards, the set G obtained is final boundary line point Group result.
CN201810817958.8A 2018-07-24 2018-07-24 Rock mass fracture identification method Expired - Fee Related CN109187548B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810817958.8A CN109187548B (en) 2018-07-24 2018-07-24 Rock mass fracture identification method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810817958.8A CN109187548B (en) 2018-07-24 2018-07-24 Rock mass fracture identification method

Publications (2)

Publication Number Publication Date
CN109187548A true CN109187548A (en) 2019-01-11
CN109187548B CN109187548B (en) 2020-04-07

Family

ID=64936610

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810817958.8A Expired - Fee Related CN109187548B (en) 2018-07-24 2018-07-24 Rock mass fracture identification method

Country Status (1)

Country Link
CN (1) CN109187548B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110080766A (en) * 2019-04-30 2019-08-02 中国矿业大学 Fully-mechanized mining working coal petrography identification device and method
CN111161252A (en) * 2019-12-31 2020-05-15 山东大学 Rock mass structure detection and dangerous stone detection system and method
CN115546113A (en) * 2022-09-15 2022-12-30 山东大学 Method and system for predicting parameters of tunnel face crack image and front three-dimensional structure
CN115761038A (en) * 2022-10-19 2023-03-07 山东大学 Tunnel face geological sketch method and system based on image spectrum technology

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102708351A (en) * 2012-05-24 2012-10-03 江南大学 Method for fast identifying Data Matrix two-dimensional bar code under complicated working condition background
US20130155109A1 (en) * 2011-11-29 2013-06-20 Pictometry International Corp. System for automatic structure footprint detection from oblique imagery
US20130223761A1 (en) * 2012-02-29 2013-08-29 Hon Hai Precision Industry Co., Ltd. Computing device and method of determining border points for measuring images of objects
CN104809689A (en) * 2015-05-15 2015-07-29 北京理工大学深圳研究院 Building point cloud model and base map aligned method based on outline
CN106225770A (en) * 2016-08-26 2016-12-14 招商局重庆交通科研设计院有限公司 Tunnel tunnel face geology multidimensional digitized record recognition methods and system
WO2017103982A1 (en) * 2015-12-14 2017-06-22 株式会社 ニコン・トリンブル Defect detection apparatus and program
CN107993244A (en) * 2017-12-27 2018-05-04 合肥市雅视智能科技有限公司 A kind of corn automatic testing method

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130155109A1 (en) * 2011-11-29 2013-06-20 Pictometry International Corp. System for automatic structure footprint detection from oblique imagery
US20130223761A1 (en) * 2012-02-29 2013-08-29 Hon Hai Precision Industry Co., Ltd. Computing device and method of determining border points for measuring images of objects
CN102708351A (en) * 2012-05-24 2012-10-03 江南大学 Method for fast identifying Data Matrix two-dimensional bar code under complicated working condition background
CN104809689A (en) * 2015-05-15 2015-07-29 北京理工大学深圳研究院 Building point cloud model and base map aligned method based on outline
WO2017103982A1 (en) * 2015-12-14 2017-06-22 株式会社 ニコン・トリンブル Defect detection apparatus and program
CN106225770A (en) * 2016-08-26 2016-12-14 招商局重庆交通科研设计院有限公司 Tunnel tunnel face geology multidimensional digitized record recognition methods and system
CN107993244A (en) * 2017-12-27 2018-05-04 合肥市雅视智能科技有限公司 A kind of corn automatic testing method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
冷彪等: "数字图像处理再隧道工程地质分析中的应用研究", 《铁道标准设计》 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110080766A (en) * 2019-04-30 2019-08-02 中国矿业大学 Fully-mechanized mining working coal petrography identification device and method
CN110080766B (en) * 2019-04-30 2024-03-01 中国矿业大学 Comprehensive mining working face coal-rock interface identification device and method
CN111161252A (en) * 2019-12-31 2020-05-15 山东大学 Rock mass structure detection and dangerous stone detection system and method
CN111161252B (en) * 2019-12-31 2022-03-25 山东大学 Rock mass structure detection and dangerous stone detection system and method
CN115546113A (en) * 2022-09-15 2022-12-30 山东大学 Method and system for predicting parameters of tunnel face crack image and front three-dimensional structure
CN115761038A (en) * 2022-10-19 2023-03-07 山东大学 Tunnel face geological sketch method and system based on image spectrum technology
CN115761038B (en) * 2022-10-19 2023-06-30 山东大学 Tunnel face geological sketch method and system based on image spectrum technology

Also Published As

Publication number Publication date
CN109187548B (en) 2020-04-07

Similar Documents

Publication Publication Date Title
CN107463918B (en) Lane line extraction method based on fusion of laser point cloud and image data
CN109187548A (en) A kind of rock cranny recognition methods
CN104217208B (en) Object detection method and device
CN105320923B (en) Model recognizing method and device
CN107346550B (en) It is a kind of for the three dimensional point cloud rapid registering method with colouring information
CN106991370B (en) Pedestrian retrieval method based on color and depth
CN104778721A (en) Distance measuring method of significant target in binocular image
CN107615334A (en) Object detector and object identification system
CN109933635A (en) A kind of method and device updating map data base
CN104102909B (en) Vehicle characteristics positioning and matching process based on lenticular information
CN110232693B (en) Image segmentation method combining thermodynamic diagram channel and improving U-Net
CN110378351B (en) Seal identification method and device
Wendel et al. Unsupervised facade segmentation using repetitive patterns
CN115690081A (en) Tree counting method, system, storage medium, computer equipment and terminal
CN113096121A (en) Pavement crack detection method and system based on cross fracture mechanics and image processing
CN111985507A (en) Rock three-dimensional point cloud joint trace extraction method
Zou et al. Path voting based pavement crack detection from laser range images
CN111091071A (en) Underground target detection method and system based on ground penetrating radar hyperbolic wave fitting
CN109543498A (en) A kind of method for detecting lane lines based on multitask network
CN110675442A (en) Local stereo matching method and system combined with target identification technology
Gong et al. Point cloud segmentation of 3D scattered parts sampled by RealSense
CN116051980B (en) Building identification method, system, electronic equipment and medium based on oblique photography
CN114742876B (en) Land vision stereo measurement method
Ziems et al. Multiple-model based verification of road data
CN108416346B (en) License plate character positioning method and device

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20200407

Termination date: 20210724

CF01 Termination of patent right due to non-payment of annual fee