CN109741401A - It is a kind of that On-line Measuring Method being installed under water for jacket based on what image restored - Google Patents
It is a kind of that On-line Measuring Method being installed under water for jacket based on what image restored Download PDFInfo
- Publication number
- CN109741401A CN109741401A CN201811586125.1A CN201811586125A CN109741401A CN 109741401 A CN109741401 A CN 109741401A CN 201811586125 A CN201811586125 A CN 201811586125A CN 109741401 A CN109741401 A CN 109741401A
- Authority
- CN
- China
- Prior art keywords
- image
- jacket
- underwater
- under water
- measuring method
- 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.)
- Withdrawn
Links
Abstract
On-line Measuring Method is installed under water for jacket based on what image restored the invention discloses a kind of, comprising the following steps: carry out Image Acquisition first, using video camera to underwater jacket, obtain original series image；Secondly, the original series image of above-mentioned acquisition is subjected to time domain mean filter by high-pass filter；Again, it is filtered by Gauss Wiener space deconvolution；Finally, by the Underwater Camera parameter model of calibration, solve hole in piece part away from.The present invention has the advantage that be suitble to underwater 3 D measurement, algorithm execution speed be lower than 0.5s, for three-dimensional measurement pitch-row detection mean error within 0.1mm, execute speed reduce, detection accuracy improve, to jacket under water install raising working efficiency.
Description
Technical field
The invention belongs to offshore wind farm fields, and in particular to a kind of to be mounted under water based on what image restored for jacket
Line measurement method.
Background technique
China's oceanic area is vast, and offshore wind energy resource is abundant, very big using the potential of wind-power electricity generation, with the new skill of wind-powered electricity generation
The independent development application of art, new material and new process, China's offshore wind farm enter the Large scale construction stage.Blower foundation is constructed
One of the important link of offshore wind farm construction.The foundation pattern of offshore wind turbine has gravity type foundation, pile formula basis,
In addition buoyant foundation technology is also in research and development.Pile formula basis includes single-pile foundation again, more pile concrete cushion cap foundations and
More pile leader frame bases.
Jacket installation include underwater pile driving, tubular pole positioning, the impurity inside tubular pole remove, underwater grouting and blower tower
Pedestal leveling etc., underwater pile driving, positioning and grouting are high for subaqueous survey precision prescribed, otherwise will will affect jacket and obtain
Attention problems progress.It is measured to solve the above problems, generalling use underwater automatic operation processing using machine vision, but
It is since refraction, the fluctuation of the water surface etc. of the water surface influence, these use machine vision subsidiary to position in automation procedure
New challenge is brought, the quality of image detection has been seriously affected.
A kind of underwater measuring method of view-based access control model of patent No. CN105698767A, using establish accurately under water at
As the Accurate Calibration of model realization camera, the three-dimensional of underwater two-dimensional measurement and binocular is realized using camera calibration parameter and is surveyed
Amount, measuring mean error for underwater two-dimension by experimental verification this method is 0.0612mm, and average for three-dimensional measurement
Error result is relatively larger than two-dimensional measurement, for less than 0.2mm, which is relatively suitble to underwater two-dimension measurement, for three-dimensional measurement,
The measurement accuracy is unable to satisfy the required precision of jacket underwater pile driving, positioning and grouting, and the measurement of the patent is spent
Time is longer.
Summary of the invention
In view of the above-mentioned deficiencies in the prior art, it is an object of the present invention to now provide it is a kind of based on image restore for jacket
Underwater installation On-line Measuring Method, which not only can effectively restore underwater disturbance image, but also can be realized view
Feel on-line measurement, algorithm execution speed is lower than 0.5s, and jacket pitch-row, which is cut, surveys mean error less than 0.1mm.
In order to solve the above technical problems, the technical solution adopted by the present invention are as follows: a kind of to be used for conduit based on what image restored
Frame installs On-line Measuring Method under water, comprising the following steps: Image Acquisition is carried out first, using video camera to underwater jacket,
Obtain original series image；Secondly, the original series image of above-mentioned acquisition is subjected to time domain mean filter by high-pass filter；
Again, it is filtered by Gauss Wiener space deconvolution；Finally, by the Underwater Camera parameter model of calibration, hole in piece part is solved
Away from.
Further, the specific steps are as follows:
A, Image Acquisition is carried out to underwater jacket using video camera, establishes Disturbance Model, enables the picture of corresponding points in image
Numerical value is X (x, y), and I (X) indicates the practical image with disturbance, Itrue(X) it indicates not by the ideal conditions of any influence of noise
Under original image, uxIndicate noise of the two-dimensional random vector at picture point X, h(u)Indicate the probability density letter of noise at this
Number, then disturbing degradation model may be expressed as:
I (X)=I (X+ux) (1)
The wherein convolution mode of formula (1) are as follows:
B, the lesser sequence image of deformation is filtered out, the original series image of acquisition is subjected to time domain by high-pass filter
Mean filter, the result handled using time domain mean filter are calculated current n frame image PSNR characteristic value and carry out height as reference picture
Pass filter, PSNR indicate Y-PSNR, MSE indicate filtering mean square error, α (i, j) andRespectively refer to
Image is the total pixel number of image with corresponding gray value in acquisition image, M × N, withIndicate binary system shared by a pixel
Digit, then αmax=2l- 1,
Wherein
C, using the space Wiener deconvolution based on Gaussian Profile, deconvolution recovery is carried out to disturbance degraded image, is obtained
Original clear image can regard the I (X) that the time domain mean filter after high-pass filtering in step B obtains as:
Wherein b (X) is the additive noise at time domain average module, therefore designs Wiener filter Wσ(f) original is obtained
Beginning location drawing picture Itrue:
Wherein f is two-dimensional frequency domain vector, and H (f) is the Fourier transformation for disturbing probability density function, StrueIt (f) is Itrue
Power spectral density, correspondingly, Sb(f) be additive noise b (X) power spectral density, Wiener filter can usually simplify are as follows:
Wherein NSR (n) is the ratio of the residual noise and signal at time domain average module output, then H (f) is given by following formula
Out:
D, the position using the template matching positioning underwater installation position of jacket in the picture, utilizes the underwater installation position
Feature holes measure the pitch-row of underwater installation position in conjunction with calibrated monocular-camera model parameter.
Further, video camera is 300,000 pixel CCD camera of Sony TELI, and camera calibration plate is standard 10mm chess
Disk lattice, scaling method use Zhang Shi standardization.
Further, the period of PSNR high-pass filtering is 50 frames in step B, and filtering threshold is set as 15 frames, i.e., every 50 frame
Original sequence filters out 15 high frame images of PSNR value.
Further, NSR (n) and σ is set all in accordance with experience in step C.
Further, the pitch-row that underwater installation position is measured in step D calculates progress as figuring method using sub-.
Beneficial effects of the present invention are as follows:
The present invention proposes that a kind of jacket that is used for restored based on image is mounted on line method under water, which can both have
Effect restores underwater disturbance image, and is able to achieve vision on-line measurement, is especially suitable for underwater 3 D measurement, algorithm is held
Scanning frequency degree be lower than 0.5s, for three-dimensional measurement pitch-row detection mean error within 0.1mm, execute speed reduce, detection
Precision improves, and installs under water to jacket and improves working efficiency.
Detailed description of the invention:
The following examples can make professional and technical personnel that the present invention be more fully understood, but therefore not send out this
It is bright to be limited among the embodiment described range.
Fig. 1 is underwater jacket detection schematic diagram of the invention.
Fig. 2 is original sequence figure of the invention.
Fig. 3 is space deconvolution image of the invention.
Fig. 4 is that canonical measure block pitch-row of the invention measures figure.
Fig. 5 is the mean error schematic diagram measured in time-series image experiment of the invention.
For time series of the invention, he thinks the algorithm execution speed measured in experiment to Fig. 6.
Specific embodiment
Embodiments of the present invention are illustrated by particular specific embodiment below, those skilled in the art can be by this explanation
Content disclosed by book is understood other advantages and efficacy of the present invention easily.
It is a kind of as shown in figure 1 that On-line Measuring Method is installed under water for jacket based on what image restored including following
Step: Image Acquisition is carried out first, using video camera to underwater jacket, obtains original series image；Secondly, by above-mentioned acquisition
Original series image pass through high-pass filter carry out time domain mean filter；Again, it is filtered by Gauss Wiener space deconvolution；
Finally, by the Underwater Camera parameter model of calibration, solve hole in piece part away from.
Specific step is as follows:
A, Image Acquisition is carried out to underwater jacket using video camera, establishes Disturbance Model, enables the picture of corresponding points in image
Numerical value is X (x, y), and I (X) indicates the practical image with disturbance, Itrue(X) it indicates not by the ideal conditions of any influence of noise
Under original image, uxIndicate noise of the two-dimensional random vector at picture point X, h(u)Indicate the probability density letter of noise at this
Number, then disturbing degradation model may be expressed as:
I (X)=I (X+ux) (1)
The wherein convolution mode of formula (1) are as follows:
B, the lesser sequence image of deformation is filtered out, the original series image of acquisition is subjected to time domain by high-pass filter
Mean filter, the result handled using time domain mean filter are calculated current n frame image PSNR characteristic value and carry out height as reference picture
Pass filter, PSNR indicate Y-PSNR, MSE indicate filtering mean square error, α (i, j) andRespectively refer to
Image is the total pixel number of image with corresponding gray value in acquisition image, M × N, withIndicate binary system shared by a pixel
Digit, then αmax=2l- 1,
Wherein
C, using the space Wiener deconvolution based on Gaussian Profile, as shown in figure 3, carrying out deconvolution to disturbance degraded image
It restores, obtains original clear image, the I (X) that the time domain mean filter after high-pass filtering in step B obtains can be regarded as:
Wherein b (X) is the additive noise at time domain average module, therefore designs Wiener filter Wσ(f) original is obtained
Beginning location drawing picture Itrue:
Wherein f is two-dimensional frequency domain vector, and H (f) is the Fourier transformation for disturbing probability density function, StrueIt (f) is Itrue
Power spectral density, correspondingly, Sb(f) be additive noise b (X) power spectral density, Wiener filter can usually simplify are as follows:
Wherein NSR (n) be the ratio of the residual noise and signal at time domain average module output, NSR (n) and σ all in accordance with
Experience setting, then H (f) is given by:
NSR (n) is set as steady state value 0.01 in step C；σ is selected according to turbulent flow size, usually 8pixels.
D, the position using the template matching positioning underwater installation position of jacket in the picture, utilizes the underwater installation position
Feature holes as shown in figure 4, measure the pitch-row of underwater installation position, surveyed in conjunction with calibrated monocular-camera model parameter
The pitch-row for measuring underwater installation position calculates progress as figuring method using sub-, improves measurement accuracy.
Sub-pixel edge positioning interpolation method maturity is high and is widely used in high-precision detection system, and basic principle is
The positioning accuracy of a pixel is obtained by template matching method first, the sub-pix information at edge is then obtained by interpolation.It inserts
Value method is in the nature to select suitable interpolating function come approximate gray-scale image edge area according to the local edge of target object
One-dimensional continuous illumination intensity function finds out the extreme point of the continuous function then further according to the correlation theory of sub-pixel edge, comes
The more Edge Feature Points coordinate of approaching to reality.If xkFor interpolation point, ykFor discrete function value, then interpolating function is
When carrying out sub-pixel edge detection, the point set at edge is obtained in advance first with mature edge detection operator
(xi, yi), then take at 3 points along the direction x according at the marginal point of gradient image midpoint R (i, j), respectively R (i-1, j), R
(i, j), R (i+1, j), then again using this 3 points gradient amplitude as functional value, with (xi- w), xi, (xi+ w) 3 points as inserting
It is worth point, wherein w is the spacing on the direction pixel x, it substitutes into above formula f (x),
It can obtain:
It is then sub-pixel edge according to the extreme point that interpolation method principle can obtain the interpolating function, thus to the interpolating function
It differentiates, that is, passes through solutionThe value that sub-pixel edge point x can be obtained is
The ordinate value of the sub-pixel edge point on the direction y can similarly be obtained.
Video camera is 300,000 pixel CCD camera of Sony TELI, and camera calibration plate is standard 10mm gridiron pattern, calibration
Method uses Zhang Shi standardization.
Test result:
The period of PSNR high-pass filtering is 50 frames in step B, and filtering threshold is set as 15 frames, i.e., every 50 frame original image sequence
Column, filter out 15 high frame images of PSNR value, then the image sequence after high-pass filtering does mean filter processing, then passes through sky
Between deconvolution obtain restored image, examine and approve segmentation through template and carry out binaryzation, there are two standards in the image after image restoration
Measuring block, each canonical measure block have 6 measured holes away from (being 20mm), 12 measured holes away from measured value it is as shown in the table:
| Calibrated bolck pitch-row | As a result (mm) | Error (mm) |
| L1 | 20.2187 | 0.2187 |
| L2 | 20.0473 | 0.0473 |
| L3 | 20.0405 | 0.0405 |
| L4 | 19.8619 | -0.1381 |
| L5 | 20.0822 | 0.0822 |
| L6 | 20.062 | 0.062 |
| L7 | 20.0276 | 0.0276 |
| L8 | 20.0362 | 0.0362 |
| L9 | 20.0261 | 0.0261 |
| L10 | 19.862 | -0.138 |
| L11 | 19.8708 | -0.1292 |
| L12 | 20.1668 | 0.1668 |
It is learnt from upper table, the three-dimensional mean error tested is 0.0252mm, mean square deviation 0.1129mm, compared to special
The three-dimensional mean error that benefit CN105698767A is tested is less than 0.2mm, therefore, using the On-line Measuring Method of this patent
Obtained from jacket installation precision be greater than the obtained precision of patent No. CN105698767A measurement method.
5 Error Graph continuously measured for 1300 frame image sequences in such as figure, Error Graph show mean error 0.1mm with
It is interior.As a result in continuous measurement process, algorithm keeps higher measurement accuracy on surface.
If Fig. 6 is shown as the algorithm execution speed figure that 1300 frame image sequences continuously measure, experiment is with computer CPU
Intel i3-2350M, core frequency 2.3GHz inside save as binary channels DDR3, capacity 8GB, when the algorithm being not optimised executes cost
Between show and averagely hold as in 449ms, algorithm execution speed is lower than 0.5s, reach the needs of on-line measurement.
Above-described embodiment is presently preferred embodiments of the present invention, is not a limitation on the technical scheme of the present invention, as long as
Without the technical solution that creative work can be realized on the basis of the above embodiments, it is regarded as falling into the invention patent
Rights protection scope in.
Claims (6)
1. a kind of install On-line Measuring Method for jacket based on what image restored under water, which is characterized in that including following step
It is rapid: to carry out Image Acquisition first, using video camera to underwater jacket, obtain original series image；Secondly, by above-mentioned acquisition
Original series image carries out time domain mean filter by high-pass filter；Again, it is filtered by Gauss Wiener space deconvolution；Most
Afterwards, by the Underwater Camera parameter model of calibration, solve hole in piece part away from.
2. a kind of according to claim 1 install On-line Measuring Method for jacket based on what image restored under water, special
Sign is, the specific steps are as follows:
A, Image Acquisition is carried out to underwater jacket using video camera, establishes Disturbance Model, enable the picture numerical value of corresponding points in image
The practical image with disturbance, I are indicated for X (x, y), I (X)true(X) it indicates not under by the ideal conditions of any influence of noise
Original image, uxIndicate that noise of the two-dimensional random vector at picture point X, h (u) indicate the probability density function of noise at this,
Then disturbing degradation model may be expressed as:
I (X)=I (X+ux) (1)
The wherein convolution mode of formula (1) are as follows:
B, the lesser sequence image of deformation is filtered out, the original series image of acquisition is subjected to time domain mean value by high-pass filter
Filtering, the result handled using time domain mean filter are calculated current n frame image PSNR characteristic value and carry out high pass filter as reference picture
Wave, PSNR indicate Y-PSNR, MSE indicate filtering mean square error, α (i, j) andRespectively reference picture
It is the total pixel number of image with corresponding gray value, M × N in acquisition image, binary digit shared by a pixel is indicated with l
It counts, then αmax=2l- 1,
Wherein
C, using the space Wiener deconvolution based on Gaussian Profile, deconvolution recovery is carried out to disturbance degraded image, is obtained original
Clear image can regard the I (X) that the time domain mean filter after high-pass filtering in step B obtains as:
Wherein b (X) is the additive noise at time domain average module, therefore designs Wiener filter Wσ(f) home position is obtained
Image Itrue:
Wherein f is two-dimensional frequency domain vector, and H (f) is the Fourier transformation for disturbing probability density function, StrueIt (f) is ItrueFunction
Rate spectrum density, correspondingly, Sb(f) be additive noise b (X) power spectral density, Wiener filter can usually simplify are as follows:
Wherein NSR (n) is the ratio of the residual noise and signal at time domain average module output, then H (f) is given by:
D, the position using the template matching positioning underwater installation position of jacket in the picture, utilizes the spy of the underwater installation position
It levies hole and measures the pitch-row of underwater installation position in conjunction with calibrated monocular-camera model parameter.
3. a kind of jacket that is used for restored based on image according to claim 1 or claim 2 installs On-line Measuring Method under water,
It is characterized in that, the video camera is 300,000 pixel CCD camera of Sony TELI, and camera calibration plate is standard 10mm chessboard
Lattice, scaling method use Zhang Shi standardization.
4. a kind of according to claim 2 install On-line Measuring Method for jacket based on what image restored under water, special
Sign is that the period of PSNR high-pass filtering is 50 frames in the step B, and filtering threshold is set as 15 frames, i.e., every 50 frame original graph
As sequence, 15 high frame images of PSNR value are filtered out.
5. a kind of according to claim 2 install On-line Measuring Method for jacket based on what image restored under water, special
Sign is that NSR (n) is set as steady state value 0.01 in the step C；σ is selected according to turbulent flow size, usually 8pixels.
6. a kind of according to claim 2 install On-line Measuring Method for jacket based on what image restored under water, special
Sign is that the pitch-row that underwater installation position is measured in the step D calculates progress as figuring method using sub-.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811586125.1A CN109741401A (en) | 2018-12-25 | 2018-12-25 | It is a kind of that On-line Measuring Method being installed under water for jacket based on what image restored |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811586125.1A CN109741401A (en) | 2018-12-25 | 2018-12-25 | It is a kind of that On-line Measuring Method being installed under water for jacket based on what image restored |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109741401A true CN109741401A (en) | 2019-05-10 |
Family
ID=66359697
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811586125.1A Withdrawn CN109741401A (en) | 2018-12-25 | 2018-12-25 | It is a kind of that On-line Measuring Method being installed under water for jacket based on what image restored |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109741401A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110847209A (en) * | 2019-10-22 | 2020-02-28 | 浙江大学 | Underwater auxiliary positioning device for offshore wind power jacket foundation hoisting by pile-first method and installation method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1447953A (en) * | 2000-08-29 | 2003-10-08 | 科学应用国际公司 | System and method for adaptive filtering |
| CN107886050A (en) * | 2017-10-16 | 2018-04-06 | 电子科技大学 | Utilize time-frequency characteristics and the Underwater targets recognition of random forest |
-
2018
- 2018-12-25 CN CN201811586125.1A patent/CN109741401A/en not_active Withdrawn
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1447953A (en) * | 2000-08-29 | 2003-10-08 | 科学应用国际公司 | System and method for adaptive filtering |
| CN107886050A (en) * | 2017-10-16 | 2018-04-06 | 电子科技大学 | Utilize time-frequency characteristics and the Underwater targets recognition of random forest |
Non-Patent Citations (1)
| Title |
|---|
| 朱军等: "一种基于图像恢复的水下零件在线测量方法", 《机电一体化》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110847209A (en) * | 2019-10-22 | 2020-02-28 | 浙江大学 | Underwater auxiliary positioning device for offshore wind power jacket foundation hoisting by pile-first method and installation method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109443307B (en) | System and method for measuring settlement and inclination angle of transmission tower based on optical measurement | |
| CN112766274B (en) | Water gauge image water level automatic reading method and system based on Mask RCNN algorithm | |
| CN109145830B (en) | Intelligent water gauge identification method | |
| Graham et al. | Automated sizing of coarse-grained sediments: image-processing procedures | |
| CN107292310B (en) | Visual positioning and automatic reading method for circular pointer type dial plate | |
| CN107274393B (en) | Monocrystaline silicon solar cell piece detection method of surface flaw based on grid line detection | |
| CN104156629B (en) | A kind of pathfinder image inverting wind direction of ocean surface method based on relative detector calibration | |
| CN107631782B (en) | Water level detection method based on Harris angular point detection | |
| CN111062978B (en) | Texture recognition method for spatio-temporal image flow measurement based on frequency domain filtering technology | |
| CN112767359B (en) | Method and system for detecting corner points of steel plate under complex background | |
| CN114743115A (en) | Shelter recognition method and fault early warning system for photovoltaic electronic equipment | |
| CN116452594A (en) | Visualized monitoring and early warning method and system for power transmission line state | |
| CN108665468B (en) | Device and method for extracting tangent tower insulator string | |
| CN104680488A (en) | Method for performing satellite-borne SAR (Synthetic Aperture Radar) image ortho-rectification based on DSM (Digital Surface Model) of median filtering | |
| CN116844147A (en) | Pointer instrument identification and abnormal alarm method based on deep learning | |
| CN102155933B (en) | Method for measuring galloping of transmission conductor on the basis of video difference analysis | |
| CN113899349B (en) | Sea wave parameter detection method, equipment and storage medium | |
| CN103413319A (en) | Industrial camera parameter on-site calibration method | |
| CN113593040B (en) | Method and system for identifying abnormal vegetation growth diseases of dike revetment | |
| CN109741401A (en) | It is a kind of that On-line Measuring Method being installed under water for jacket based on what image restored | |
| CN112085725B (en) | Residual film residual quantity detection method and early warning system based on heuristic iterative algorithm | |
| CN111476761B (en) | Visibility measurement method and system based on system identification | |
| CN114863258B (en) | Method for detecting small target based on visual angle conversion in sea-sky-line scene | |
| CN102183239A (en) | Method for measuring windage yaw of transmission wire based on video variance analysis | |
| CN116363522A (en) | Coastal zone reclamation sea change remote sensing monitoring method based on deep learning |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20190510 |
|
| WW01 | Invention patent application withdrawn after publication |