CN106651843A - Image processing method for tunnel water seepage detection - Google Patents
Image processing method for tunnel water seepage detection Download PDFInfo
- Publication number
- CN106651843A CN106651843A CN201611157221.5A CN201611157221A CN106651843A CN 106651843 A CN106651843 A CN 106651843A CN 201611157221 A CN201611157221 A CN 201611157221A CN 106651843 A CN106651843 A CN 106651843A
- Authority
- CN
- China
- Prior art keywords
- region
- image
- infiltration
- value
- black
- 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
Classifications
-
- 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
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
Landscapes
- Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an image processing method for tunnel water seepage detection, and belongs to the technical field of image processing. The method is characterized by comprising the following steps of when a black water seepage region is detected, firstly performing Gaussian filtering on an original image, secondly removing a region in which a rule line in an obtained result is located, thirdly performing mean value-based black region detection and brightness value and contrast ratio-based feature measurement on the image and removing an ineffective region, and finally performing transverse and longitudinal measurement on the whole image to obtain a tunnel image marked with the black water seepage region; and when a reflected water seepage region is detected, firstly performing salient region detection on the image subjected to the Gaussian filtering, secondly performing threshold segmentation according to a gradient value and calculating a coincided part of the salient region and a region with a relatively large gradient, and finally removing a region containing relatively few salient pixels from a coincided region to obtain a tunnel image marked with the reflected water seepage region. Compared with an infrared thermal imaging method and a laser scanning method, the scheme has the advantages that the accuracy is high and the detection cost is reduced.
Description
Technical field
The invention belongs to technical field of image processing, and in particular to a kind of image processing method of water seepage of tunnel detection.
Background technology
In recent years, China's track traffic quickly grows, and particularly Tunnel Engineering is more and more adopted in track traffic
With, but due to the concrete shrinkage and creep used in Tunnel Engineering, weathering, and some geological disasters or human factor
The structure destruction for causing so that the maintenance problem of a large amount of tunnels facility is displayed, and the Tunnel Engineering in track traffic is made in advance
Enter curing period.If can predict in tunnel occur potential safety hazard or at the initial stage that facility breaks down if it is taken
Corresponding measure, it will substantially reduce maintenance costs, improves the safety index of circuit.
Water seepage of tunnel becomes the key factor for affecting tunnel structure intensity in all kinds of tunnel defects.Water seepage of tunnel is to tunnel
Other diseases of road have direct or indirect impact, are to be related to one of most commonly used tunnel defect.Present on tunnel structure
Defect easily causes the generation of pore gas flow, and the long term for seeping water can aggravate the destruction of tunnel structure, water seepage of tunnel reality
The concentrated expression of the various diseases in border Shang Shi tunnels.It is to realize effectively in advance so strengthening the research of tunnel-liner infiltration detection method
The necessary ways of anti-tunnel defect potential hazard, make a prediction to potential safety hazard that may be present.
Generally using the method for artificial range estimation, have that efficiency is low, subjectivity is strong etc. lacks existing water seepage of tunnel conventional detection
Point, therefore over the years people put forth effort on always the research of the Fast nondestructive evaluation technology of water seepage of tunnel.At present using relatively broad
Method have:Infrared thermography is the infiltration automatic testing method being most widely used at present, its thing according to different temperatures
The infrared emanation of body is different to realize infiltration detection.Another kind of laser scanning nondestructive determination is also being developed rapidly, and it can be to tunnel
Carry out it is comprehensive retouch detection, so as to record infiltration size and position, with retouching the features such as speed is fast, certainty of measurement is high,
But said method haves the shortcomings that to spend high cost.
The content of the invention
It is an object of the invention to provide a kind of image processing method of water seepage of tunnel detection, can effectively utilize at image
The method of reason detects and marks black in tunnel image to seep water region and reflective infiltration region, and overcomes in prior art and deposit
Shortcoming.
The present invention is achieved in that it is characterized in that implementation steps are as follows:
The first step:Black infiltration region detection, comprises the following steps:
First, gaussian filtering:Piece image D is read in, its resolution ratio is p × q, and D is divided into into 3000 × 3000 little image
Jing gaussian filterings obtain imm after im;
2nd, detect and remove the region that regular lines are located in imm and obtain valid:Rule is detected and removed to image imm
The region imline_pipe that the region imline_vh and tubular articles that near linear object is located is located obtains region valid:
The region imline_vh that the near linear object of the rule is located must simultaneously meet following feature:
1) long axis length in the region must be more than 20 with the ratio axis of minor axis length, to guarantee that the region is approximately straight
Line;
If 2) the region major axis is judged to vertical line with the x-axis angle of cut more than 85 degree or less than -85 degree;
If 3) the region major axis is judged to parallel lines with the x-axis angle of cut more than -5 degree and less than 5 degree;
The region imline_pipe that the tube is located must simultaneously meet following feature:
1) the region long axis length must be more than 90 with the ratio axis of minor axis length, to guarantee that the region is approximately straight line;
2) ratio r atio of two line lengths must meet 0.8 < ratio < 1.2, to guarantee the approximate phase of two line lengths
Deng;
3) think on the lines of equal length at two, by two different positions the distance of this two lines is obtained, by institute
Two distances for obtaining do product and obtain dis, and dis less than the 1/5 of testing image size minimum of a value;
3rd, valid is done and valid3 is detected to obtain based on the black region of average:
First the noise such as region such as metope background in valid is removed according to average and variance and obtain valid1:1. basis
Area pixel average enters row threshold division to image:Remove more than the part of threshold value in region, threshold value k is determined by formula (1),
Mean2 (valid) refers to take average to area pixel in formula;
2. according to the variance of current line, noise region is further removed:If the variance of current line pixel value is more than 4,
Remove current line;
Secondly, line card region is removed in valid1 and obtains valid2:Because line card is generally internal shinny, surrounding is sent out
Secretly, therefore, if occurring cavity, i.e. this less shinny region inside the black region for detecting, then it is assumed that the region
It is line card region, rather than region of seeping water, remove it;
Finally, remove in valid2 and there is the region of symmetrical property and part brighter areas to obtain valid3:Definition is treated
It is brighter areas more than the part of the regional luminance average+20 to survey brightness in region, then when brighter areas area is more than current
The 1/4 of region area, then it is assumed that the region is not nigrescence color infiltration region and removes it;
4th, based on the measurement to brightness value and contrast, remove in valid3 that brightness is excessive and luminance mean value and surrounding
The less region of environment difference obtains valid4:1. row threshold division is entered to image according to image pixel average:Definition is more than threshold value
Regional luminance it is excessive, be not black infiltration region, remove it, threshold value k by formula (2) determine,
Mean2 (valid4) is referred to image averaging in formula;
2. the contrast in region to be measured and neighboring pixel is measured, because the dark infiltration region of black often has with surrounding environment
Larger luminance difference, it is believed that the region with surrounding environment luminance difference less than 3.5 is not black infiltration region, removes it;
5th, cross measure is carried out to the big figure of view picture and obtains im_label with longitudinal direction measurement:Horizontal stroke is taken in measured region
To the maximum with longitudinal measured value:
First carry out cross measure:In measured region, diverse location in region is measured according to the abscissa in region
Width value, with the width that the maximum of width value represents the region;Longitudinal measurement is carried out again:If area coordinate position of centre of gravity
The difference of maxima and minima is less than 700, then do not carry out longitudinal measurement;Otherwise, according to the size of ordinate average in region to be measured,
Region is classified with dichotomy, the gathering for a class less than 700 by the difference of all kinds of ordinates, finds out the maximum class of length
Can ensure that longitudinal region is most long, so that it is determined that the distribution in the region;
6th, the black infiltration region that output is detected:More than Jing operations remove all kinds of inactive areas, obtain black infiltration area
Domain, is marked and is exported and obtained im_out;
Second step:Reflective infiltration region detection, comprises the following steps:
First, the salient region based on average is done to imm to detect, finds the big region im_salient of brightness:First by imm
It is converted into double types and obtains region k to be measured, then k is asked for be worth to lm, finally to each location of pixels in the lm of region
Be for further processing by formula (3) and formula (4), then Jing small areas region remove etc. operation obtain salient region im_
salient:
Sm (x, y)=[l (x, y)-lm]2 (3)
X=1,2,3 in formula ..., 3000;Y=1,2,3 ..., 3000
Sm (x, y) is the variance of corresponding pixel points
Smini (x, y) is the business of the variance with the Local Deviation maximum of each pixel
2nd, row threshold division is entered according to region gradient to imm and obtains bw:Formula is used to each pixel imm (x, y) in imm
(5) gradient is asked for, and row threshold division is entered to image according to Grad, retain region of the gradient more than 20, then expanded operation
Process with hole-filling etc. and obtain bianry image bw:
X=1,2,3 in formula ..., 3000;Y=1,2,3 ..., 3000, wxAnd hyTo be respectively pixel imm (x, y)
Widthwise edge edge value and longitudinal edge edge value;
3rd, the overlapping region im_coin of salient region im_salient and gradient large area bw is retained:Define these
Overlapping region is reflective infiltration region;
4th, remove in im_coin and obtain im_coin1 comprising the less region of conspicuousness pixel:Remaining comprising aobvious
In the gradient large area in work property region, if salient region occupied area is less than 1/64, the Ze Gai areas of the region gross area
Domain is not reflective infiltration region, is removed it;
5th, the reflective infiltration region for detecting is exported:More than Jing operations remove all kinds of inactive areas, obtain reflective infiltration area
Domain, is marked and is exported Doutput。
Advantages of the present invention and good effect are:
1st, the method that the present invention is adopted is compared with conventional water seepage of tunnel detection method, right without manually going to scene in person
The body of wall in tunnel is observed, and use manpower and material resources sparingly resource, it also avoid the danger of lining cutting flake-off block.
2nd, the method is compared with infrared thermography with laser scanning method, not only ensure that the existing high precision of both approaches
Rate, is greatly reduced the cost needed for detection again, reduces unnecessary expenditures.
3rd, the different qualities difference that the present invention is showed in the picture using black infiltration region and reflective infiltration region
Two class regions are detected, all possible infiltration region is detected, it is ensured that the validity of method proposed by the invention.
Description of the drawings
Fig. 1 is the enforcement block diagram of the present invention;
Fig. 2 is the original image of input;
Fig. 3 be input original tunnel image it is filtered after result;
Fig. 4 is image of the little figure Jing after regular lines region detection in Fig. 3;
Result after black infiltration region detections of the Fig. 5 based on average;
Fig. 6 is based on the result after brightness and contrast's detection;
Fig. 7 be to the big figure of view picture Jing cross measure and longitudinal direction measurement after gained black infiltration region;
Salient region detection figures of the Fig. 8 based on average;
Fig. 9 is based on the Threshold segmentation result of Grad;
The overlapping region result of Figure 10 salient regions and gradient large area;
Figure 11 is removed in Fig. 9 comprising the area results that conspicuousness pixel is less;
Figure 12 is the tunnel image in mark infiltration region.
Specific embodiment
To water seepage of tunnel region detection scheme proposed by the present invention, we have done preliminary test experiments.Using a tunnel
Image seep water as input picture in road.Make detection process using Asus's notebook computer, notebook parameter is:Intel(R),
Core (TM) i5CPU, 3210 ,@2.5GHz, 4.00GB internal memories.Software platform is MatlabR2014a, uses Matlab Programming with Pascal Language
Realize water seepage of tunnel region detection scheme.
Fig. 1 gives flow chart of the present invention, it is characterised in that implementation steps are as follows:
The first step:Black infiltration region detection, comprises the following steps:
First, gaussian filtering:Read in piece image D be Fig. 2, its resolution ratio be p × q, by D be divided into 3000 × 3000 it is little
To obtain imm as shown in Figure 3 for Jing gaussian filterings after image im;
2nd, detect and remove the region that regular lines are located in imm and obtain valid:Rule is detected and removed to image imm
The region imline_pipe that the region imline_vh and tubular articles that near linear object is located is located obtains region valid,
As shown in Figure 4:
The region imline_vh that the near linear object of the rule is located must simultaneously meet following feature:
1) long axis length in the region must be more than 20 with the ratio axis of minor axis length, to guarantee that the region is approximately straight
Line;
If 2) the region major axis is judged to vertical line with the x-axis angle of cut more than 85 degree or less than -85 degree;
If 3) the region major axis is judged to parallel lines with the x-axis angle of cut more than -5 degree and less than 5 degree;
The region imline_pipe that the tube is located must simultaneously meet following feature:
1) the region long axis length must be more than 90 with the ratio axis of minor axis length, to guarantee that the region is approximately straight line;
2) ratio r atio of two line lengths must meet 0.8 < ratio < 1.2, to guarantee the approximate phase of two line lengths
Deng;
3) think on the lines of equal length at two, by two different positions the distance of this two lines is obtained, by institute
Two distances for obtaining do product and obtain dis, and dis less than the 1/5 of testing image size minimum of a value;
3rd, to obtain valid3 as shown in Figure 5 for the black infiltration region detection for doing to valid based on average:
First the noise such as region such as metope background in valid is removed according to average and variance and obtain valid1:1. basis
Area pixel average enters row threshold division to image:Remove more than the part of threshold value in region, threshold value k is determined by formula (1),
Mean2 (valid) refers to take average to area pixel in formula;
2. according to the variance of current line, noise region is further removed:If the variance of current line pixel value is more than 4,
Remove current line;
Secondly, line card region is removed in valid1 and obtains valid2:Because line card is generally internal shinny, surrounding is sent out
Secretly, therefore, if occurring cavity, i.e. this less shinny region inside the black region for detecting, then it is assumed that the region
It is line card region, rather than region of seeping water, remove it;
Finally, remove in valid2 and there is the region of symmetrical property and part brighter areas to obtain valid3:Definition is treated
It is brighter areas more than the part of the regional luminance average+20 to survey brightness in region, then when brighter areas area is more than current
The 1/4 of region area, then it is assumed that the region is not nigrescence color infiltration region and removes it;
4th, based on the measurement to brightness value and contrast, remove in valid3 that brightness is excessive and luminance mean value and surrounding
It is as shown in Figure 6 that the less region of environment difference obtains valid4:1. row threshold division is entered to image according to image pixel average:Definition
It is excessive more than the regional luminance of threshold value, it is not black infiltration region, remove it, threshold value k is determined by formula (2),
Mean2 (valid4) is referred to image averaging in formula;
2. the contrast in region to be measured and neighboring pixel is measured, because the dark infiltration region of black often has with surrounding environment
Larger luminance difference, it is believed that the region with surrounding environment luminance difference less than 3.5 is not black infiltration region, removes it;
5th, the cross measure and longitudinal direction measurement to the big figure of view picture obtains im_label:Take in measured region laterally
With the maximum of longitudinal measured value:
First carry out cross measure:In measured region, diverse location in region is measured according to the abscissa in region
Width value, with the width that the maximum of width value represents the region;Longitudinal measurement is carried out again:If area coordinate position of centre of gravity
The difference of maxima and minima is less than 700, then do not carry out longitudinal measurement;Otherwise, according to the size of ordinate average in region to be measured,
Region is classified with dichotomy, the gathering for a class less than 700 by the difference of all kinds of ordinates, finds out the maximum class of length
Can ensure that longitudinal region is most long, so that it is determined that the distribution in the region;
6th, the black infiltration region that output is detected:More than Jing operations remove all kinds of inactive areas, obtain black infiltration area
Domain, is marked and is exported that to obtain im_out as shown in Figure 7;
Second step:Reflective infiltration region detection, comprises the following steps:
First, the salient region based on average is done to imm to detect, finds the big region im_salient such as Fig. 8 institutes of brightness
Show:First imm is converted into into double types and obtains region k to be measured, then k is asked for be worth to lm, finally to every in the lm of region
Individual location of pixels is for further processing by formula (3) and formula (4), then Jing small areas region remove etc. operation obtain conspicuousness area
Domain im_salient;
Sm (x, y)=[l (x, y)-lm]2 (3)
X=1,2,3 in formula ..., 3000;Y=1,2,3 ..., 3000
Sm (x, y) is the variance of corresponding pixel points
Smini (x, y) is the business of the variance with the Local Deviation maximum of each pixel
2nd, entering row threshold division according to region gradient to imm, to obtain bw as shown in Figure 9:To each pixel imm in imm (x,
Y) gradient is asked for formula (5), and row threshold division is entered to image according to Grad, retain region of the gradient more than 20, then Jing
Expansive working and hole-filling etc. are processed and obtain bianry image bw;
Wherein x=1,2,3 ..., 3000;Y=1,2,3 ..., 3000, wxAnd hyTo be respectively pixel imm (x, y)
Widthwise edge edge value and longitudinal edge edge value;
3rd, overlapping region im_coin such as Figure 10 institutes of salient region im_salient and gradient large area bw are retained
Show:These overlapping regions are defined for reflective infiltration region;
4th, im_coin1 is obtained comprising the less region of conspicuousness pixel in removal im_coin as shown in figure 11:Retaining
In the gradient large area comprising salient region come, if salient region occupied area is less than the 1/ of the region gross area
64, then the region is not reflective infiltration region, is removed it;
5th, the reflective infiltration region for detecting is exported:More than Jing operations remove all kinds of inactive areas, obtain reflective infiltration area
Domain, is marked and is exported Doutput, as shown in figure 12.
In this experiment, Fig. 2 to Fig. 7 is the figure produced during black infiltration region detection, and Fig. 8 to Figure 12 is reflective oozing
The figure produced in aqua region detection process.Wherein Fig. 3 is the filtered rear resulting image of original tunnel image of input;Fig. 4
It is that the little figure that will scheme greatly to be split removes the image behind regular lines region;Fig. 5 is the black in the present invention based on average
Result after region detection;Fig. 6 is based on the result after brightness and contrast's detection;Fig. 7 is figure Jing cross measures big to view picture
With gained black infiltration region after the measurement of longitudinal direction.Fig. 8 is based on the salient region detection figure of average;Fig. 9 is based on Grad
Threshold segmentation result;Figure 10 is the overlapping region of salient region and gradient large area;Figure 11 is to remove to be included in Figure 10
The less region acquired results of conspicuousness pixel;Figure 12 is the tunnel image in mark infiltration region.
This programme repeatedly carries out expansive working, small area region during infiltration region detection is carried out to tunnel image
The process such as removal, hole-filling, to ensure the accuracy in detected infiltration region, makes result more credible.The present invention is in tunnel
Validity in the detection of road infiltration is also verified, and water seepage of tunnel Defect inspection provides a kind of new detection side for after
Method, the automatic identification for also detecting for water seepage of tunnel provides theoretical foundation.It can be seen that, water seepage of tunnel detection image processing method with
The advantages of its is quick, lossless, efficient, has good application prospect in terms of tunnel-liner infiltration generaI investigation.
Claims (1)
1. a kind of image processing method of water seepage of tunnel detection, can effectively utilize image processing method and detect and mark tunnel
Black infiltration region and reflective infiltration region in image, it is characterised in that implementation steps are as follows:
The first step:Black infiltration region detection, comprises the following steps:
First, gaussian filtering:Piece image D is read in, its resolution ratio is p × q, and D is divided into after 3000 × 3000 little image im
Jing gaussian filterings obtain imm;
2nd, detect and remove the region that regular lines are located in imm and obtain valid:The approximate of rule is detected and removed to image imm
The region imline_pipe that the region imline_vh and tubular articles that straight line object is located is located obtains region valid:
The region imline_vh that the near linear object of the rule is located must simultaneously meet following feature:
1) long axis length in the region must be more than 20 with the ratio axis of minor axis length, to guarantee that the region is approximately straight line;
If 2) the region major axis is judged to vertical line with the x-axis angle of cut more than 85 degree or less than -85 degree;
If 3) the region major axis is judged to parallel lines with the x-axis angle of cut more than -5 degree and less than 5 degree;
The region imline_pipe that the tube is located must simultaneously meet following feature:
1) the region long axis length must be more than 90 with the ratio axis of minor axis length, to guarantee that the region is approximately straight line;
2) ratio r atio of two line lengths must meet 0.8 < ratio < 1.2, to guarantee two line length approximately equals;
3) think on the lines of equal length at two, the distance of this two lines is obtained by two different positions, by gained
Two distances do product and obtain dis, and dis less than the 1/5 of testing image size minimum of a value;
3rd, valid is done and valid3 is detected to obtain based on the black region of average:
First the noise such as region such as metope background in valid is removed according to average and variance and obtain valid1:1. according to region
Pixel average enters row threshold division to image:Remove more than the part of threshold value in region, threshold value k is determined by formula (1),
Mean2 (valid) refers to take average to area pixel in formula (1);
2. according to the variance of current line, noise region is further removed:If the variance of current line pixel value is more than 4, remove
Current line;
Secondly, line card region is removed in valid1 and obtains valid2:Because line card is generally internal shinny, surrounding shades, because
This, if occurring cavity, i.e. this less shinny region inside the black region for detecting, then it is assumed that the region is line card
Region, rather than region of seeping water, remove it;
Finally, remove in valid2 and there is the region of symmetrical property and part brighter areas to obtain valid3:Define area to be measured
Brightness in domain is brighter areas more than the part of the regional luminance average+20, then when brighter areas area is more than current region
The 1/4 of area, then it is assumed that the region is not nigrescence color infiltration region and removes it;
4th, based on the measurement to brightness value and contrast, remove in valid3 that brightness is excessive and luminance mean value and surrounding environment
The less region of difference obtains valid4:1. row threshold division is entered to image according to image pixel average:Area of the definition more than threshold value
Domain brightness is excessive, is not black infiltration region, removes it, and threshold value k is determined by formula (2),
Mean2 (valid4) is referred to image averaging in formula;
2. the contrast in region to be measured and neighboring pixel is measured, because the dark infiltration region of black often has larger with surrounding environment
Luminance difference, it is believed that with surrounding environment luminance difference less than 3.5 region be not black infiltration region, remove it;
5th, cross measure is carried out to the big figure of view picture and obtains im_label with longitudinal direction measurement:Take in measured region laterally and
The maximum of longitudinal measured value:
First carry out cross measure:In measured region, the width of diverse location in region is measured according to the abscissa in region
Angle value, with the width that the maximum of width value represents the region;Longitudinal measurement is carried out again:If the maximum of area coordinate position of centre of gravity
The difference of value and minimum of a value is less than 700, then do not carry out longitudinal measurement;Otherwise, according to the size of ordinate average in region to be measured, with two
Point-score is classified to region, the gathering for a class less than 700 by the difference of all kinds of ordinates, finds out the maximum class of length
Ensure that longitudinal region is most long, so that it is determined that the distribution in the region;
6th, the black infiltration region that output is detected:More than Jing operations remove all kinds of inactive areas, obtain black infiltration region,
Marked and exported and obtained im_out;
Second step:Reflective infiltration region detection, comprises the following steps:
First, the salient region based on average is done to imm to detect, finds the big region im_salient of brightness:First imm is converted
Region k to be measured is obtained for double types, then k is asked for be worth to lm, finally to each location of pixels in the lm of region by public affairs
Formula (3) and formula (4) are for further processing, then Jing small areas region remove etc. operation obtain salient region im_salient;
Sm (x, y)=[l (x, y)-lm]2 (3)
X=1,2,3 in formula ..., 3000;Y=1,2,3 ..., 3000
Sm (x, y) is the variance of corresponding pixel points
Smini (x, y) is the business of the variance with the Local Deviation maximum of each pixel
2nd, row threshold division is entered according to region gradient to imm and obtains bw:Each pixel imm (x, y) in imm is asked with formula (5)
Gradient is taken, and row threshold division is entered to image according to Grad, retain region of the gradient more than 20, then expanded operation and cavity
The process such as fill up and obtain bianry image bw;
Wherein x=1,2,3 ..., 3000;Y=1,2,3 ..., 3000, wxAnd hyTo be respectively the horizontal of pixel imm (x, y)
Marginal value and longitudinal edge edge value;
3rd, the overlapping region im_coin of salient region im_salient and gradient large area bw is retained:Define these coincidences
Region is reflective infiltration region;
4th, remove in im_coin and obtain im_coin1 comprising the less region of conspicuousness pixel:Remaining comprising conspicuousness
In the gradient large area in region, if salient region occupied area is less than the 1/64 of the region gross area, the region is not
It is reflective infiltration region, removes it;
5th, the reflective infiltration region for detecting is exported:More than Jing operations remove all kinds of inactive areas, obtain reflective infiltration region,
Marked and exported Doutput。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611157221.5A CN106651843B (en) | 2016-12-15 | 2016-12-15 | The image processing method of water seepage of tunnel detection |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611157221.5A CN106651843B (en) | 2016-12-15 | 2016-12-15 | The image processing method of water seepage of tunnel detection |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106651843A true CN106651843A (en) | 2017-05-10 |
CN106651843B CN106651843B (en) | 2019-09-03 |
Family
ID=58822409
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611157221.5A Active CN106651843B (en) | 2016-12-15 | 2016-12-15 | The image processing method of water seepage of tunnel detection |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106651843B (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102589808A (en) * | 2012-01-16 | 2012-07-18 | 苏州临点三维科技有限公司 | Large-scale tunnel seepage point measuring method |
WO2014006234A1 (en) * | 2012-07-04 | 2014-01-09 | Repsol, S.A. | Integrated system for the early detection of hydrocarbons in an aqueous medium |
CN104048969A (en) * | 2014-06-19 | 2014-09-17 | 樊晓东 | Tunnel defect recognition method |
CN105120224A (en) * | 2015-09-09 | 2015-12-02 | 招商局重庆交通科研设计院有限公司 | Tunnel water disaster intelligent monitoring system and method based on video event detection |
-
2016
- 2016-12-15 CN CN201611157221.5A patent/CN106651843B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102589808A (en) * | 2012-01-16 | 2012-07-18 | 苏州临点三维科技有限公司 | Large-scale tunnel seepage point measuring method |
WO2014006234A1 (en) * | 2012-07-04 | 2014-01-09 | Repsol, S.A. | Integrated system for the early detection of hydrocarbons in an aqueous medium |
CN104048969A (en) * | 2014-06-19 | 2014-09-17 | 樊晓东 | Tunnel defect recognition method |
CN105120224A (en) * | 2015-09-09 | 2015-12-02 | 招商局重庆交通科研设计院有限公司 | Tunnel water disaster intelligent monitoring system and method based on video event detection |
Non-Patent Citations (3)
Title |
---|
I.M. LEE,S.W. NAM: "The study of seepage forces acting on the tunnel lining and tunnel face in shallow tunnels", 《TUNNELLING AND UNDERGROUND SPACE TECHONOLOGY》 * |
Q.M. ZHANG等: "Groundwater Disease of Tunnel Engineering Inspection By impactful detection method", 《2012 2ND INTERNATIONAL CONFERENCE ON REMOTE SENSING, ENVIRONMENT AND TRANSPORTATION ENGINEERING》 * |
豆海涛 等: "隧道渗漏水红外辐射特征模型试验及图像处理", 《岩石力学与工程学报》 * |
Also Published As
Publication number | Publication date |
---|---|
CN106651843B (en) | 2019-09-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Xu et al. | Automatic defect detection and segmentation of tunnel surface using modified Mask R-CNN | |
Lei et al. | A novel tunnel-lining crack recognition system based on digital image technology | |
Pan et al. | Automatic sewer pipe defect semantic segmentation based on improved U-Net | |
Li et al. | Integrated processing of image and GPR data for automated pothole detection | |
Tang et al. | Automatic crack detection and segmentation using a hybrid algorithm for road distress analysis | |
Yu et al. | Pavement pothole detection and severity measurement using laser imaging | |
Luo et al. | A fast adaptive crack detection algorithm based on a double-edge extraction operator of FSM | |
Sun et al. | Weighted neighborhood pixels segmentation method for automated detection of cracks on pavement surface images | |
Mathavan et al. | Pavement raveling detection and measurement from synchronized intensity and range images | |
CN102680480A (en) | Intelligent detecting method for cracks of concrete structures | |
CN105387801B (en) | A kind of subway tunnel section of jurisdiction faulting of slab ends quantity measuring method | |
CN111507971A (en) | Tunnel surface defect detection method | |
Chen et al. | A potential crack region method to detect crack using image processing of multiple thresholding | |
Dan et al. | Application of deep learning-based image recognition technology to asphalt–aggregate mixtures: Methodology | |
CN103914829B (en) | Method for detecting edge of noisy image | |
Chen et al. | An improved minimal path selection approach with new strategies for pavement crack segmentation | |
CN106504246A (en) | The image processing method of tunnel slot detection | |
Li et al. | Pixel-level detection and measurement of concrete crack using faster region-based convolutional neural network and morphological feature extraction | |
Parrany et al. | A new image processing strategy for surface crack identification in building structures under non‐uniform illumination | |
JP2008267943A (en) | Crack detecting method | |
CN109767426B (en) | Shield tunnel water leakage detection method based on image feature recognition | |
Su et al. | Fractal theory based identification model for surface crack of building structures | |
CN106651843A (en) | Image processing method for tunnel water seepage detection | |
CN113744219B (en) | Rock joint fracture overall complexity measurement analysis method based on improved fractal science | |
Kaur et al. | An Efficient Method of Number Plate Extraction from Indian Vehicles Image |
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 |