CN106529542A - Indicator diagram identification method and device - Google Patents
Indicator diagram identification method and device Download PDFInfo
- Publication number
- CN106529542A CN106529542A CN201610873368.8A CN201610873368A CN106529542A CN 106529542 A CN106529542 A CN 106529542A CN 201610873368 A CN201610873368 A CN 201610873368A CN 106529542 A CN106529542 A CN 106529542A
- Authority
- CN
- China
- Prior art keywords
- pixel
- gradient
- gray value
- numbering
- indicator card
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 75
- 238000010586 diagram Methods 0.000 title abstract 10
- 238000006243 chemical reaction Methods 0.000 claims description 10
- 238000012549 training Methods 0.000 claims description 4
- 238000009825 accumulation Methods 0.000 claims description 3
- 206010016256 fatigue Diseases 0.000 claims 1
- 238000012545 processing Methods 0.000 abstract description 4
- 230000008569 process Effects 0.000 description 12
- 230000006870 function Effects 0.000 description 9
- 238000006073 displacement reaction Methods 0.000 description 5
- 238000004519 manufacturing process Methods 0.000 description 5
- 239000003129 oil well Substances 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 229930195733 hydrocarbon Natural products 0.000 description 4
- 150000002430 hydrocarbons Chemical class 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 239000000725 suspension Substances 0.000 description 3
- 230000001186 cumulative effect Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 230000007935 neutral effect Effects 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005183 dynamical system Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000007087 memory ability Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 235000015170 shellfish Nutrition 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
- G06V10/507—Summing image-intensity values; Histogram projection analysis
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The application provides a method and a device for identifying an indicator diagram, wherein the method comprises the following steps: according to the first gray value of each pixel point in the indicator diagram, obtaining a second gray value of the pixel point through binarization processing; obtaining the gradient amplitude and the angle of the gradient direction of each pixel point according to the second gray value of the pixel point; obtaining a gradient direction histogram of the indicator diagram according to the gradient amplitude and the gradient direction angle of each pixel point; and identifying the type of the indicator diagram according to the gradient direction histogram of the indicator diagram. The method and the device for recognizing the indicator diagram type solve the technical problems of low recognition efficiency and low recognition accuracy of the existing recognition method by recognizing the indicator diagram type through the gradient direction histogram of the indicator diagram, and improve the recognition speed and accuracy; in addition, the gradient direction histogram can be directly obtained according to the graph of the indicator diagram, so that the dependence on original data is avoided, and the identification accuracy is further improved.
Description
Technical field
The present invention relates to production of hydrocarbons technical field, more particularly to a kind of indicator card recognition methods and device.
Background technology
During production of hydrocarbons, indicator card can be the closing being made up of with the relation curve that displacement is gradually changed load
Curve, can react the working condition of oil pumper.Therefore, indicator card can be a kind of important evidence of production of hydrocarbons management.Root
According to the type and other information of indicator card, the operating mode of oil pumper can be diagnosed, and then can quickly grasp the work of oil well
Make state, whether rationally analysis judges the parameter of oil well, with the working condition of oil well and oilwell parameter as foundation, can be in time
Oil well is adjusted, so as to reach the purpose for reducing loss, improving oil and gas production.
Due to indicator card during concrete production of hydrocarbons with important function, therefore, how quickly and accurately to recognize
The type of indicator card is an important problem.At present, the indicator card recognition methods for generally adopting can be generally divided into two
The individual stage：Feature extraction and classification.Wherein, feature extraction is general can be further subdivided into again：Geometrical measurers, gridding method,
Vector characteristic method, the Gray Moment tactical deployment of troops, moment characteristics and Spectrum Method.Above-mentioned several extracting methods due to the limitation of method itself, greatly
Some primary characteristic quantities can only all be extracted.But, what is obtained is identified to indicator card using these primary characteristic quantities
As a result the degree of accuracy is not often high.Simultaneously as above-mentioned several method is required for greatly being processed using the former data of indicator card,
And the method that different indicator cards is obtained is different with the mode of process preservation, correspondingly, the process to former data is needed using not
Same mode.During concrete process, a kind of mode is used uniformly across to all of former data often and is processed, so can be to recognition result
Cause error.Additionally, the Part Methods in said method, such as vector characteristic method and moment characteristics method etc., as method itself sets
The problem of meter, is vulnerable to the impact of noise, causes the reduction of the recognition result degree of accuracy.Summary situation, existing indicator card
Recognition methods has that recognition efficiency is low and the low technical problem of recognition accuracy in the specific implementation, often.
For the problems referred to above, effective solution is not yet proposed at present.
The content of the invention
A kind of indicator card recognition methods and device is embodiments provided, the gradient direction by using indicator card is straight
The type of side's figure identification indicator card, to reach present in the existing recognition methods of solution, recognition accuracy is low and recognition efficiency is low
Technical problem.
A kind of indicator card recognition methods is embodiments provided, including：
According to the first gray value of each pixel in indicator card to be identified, binary conversion treatment is carried out to the indicator card,
Obtain the second gray value of each pixel in the indicator card；
According to the second gray value of each pixel, solution obtains the ladder of the second gray value of each pixel
The angle of degree amplitude and gradient direction；
According to the angle of the gradient magnitude and gradient direction of each pixel, solution obtains the gradient of the indicator card
Direction histogram；
According to the gradient orientation histogram of the indicator card, the type of the indicator card is recognized.
In one embodiment, first gray value according to each pixel in indicator card to be identified, to described
Indicator card carries out binary conversion treatment, obtains the second gray value of each pixel in the indicator card, including：
Second gray value of first gray value in described each pixel more than or equal to the pixel of predetermined threshold value is set to
1；
Second gray value of first gray value in described each pixel less than the pixel of the predetermined threshold value is set to
0。
In one embodiment, the second gray value according to each pixel, solution obtain described each pixel
The gradient magnitude of the second gray value of point and the angle of gradient direction, including：
In such a way, it is i to number in asking for described each pixel, the gradient of the second gray value of the pixel of j
The angle of amplitude and gradient direction：
It is i-1 according to numbering, the second gray value of the pixel of j, numbering are i+1, the second gray value of the pixel of j,
Numbering is i, and the second gray value of the pixel of j-1 and numbering are i, and the second gray value of the pixel of j+1, solution obtain described
Numbering is i, the horizontal direction gradient component and vertical gradient component of the second gray value of the pixel of j；
It is i, the horizontal direction gradient component of the second gray value of the pixel of j and vertical gradient according to the numbering
Component, it is i, the gradient magnitude of the second gray value of the pixel of j and the angle of gradient direction that solution obtains numbering；
Wherein, i be 1 to the integer between n, j be 1 to the integer between m, n × m be in the indicator card pixel
Number.
In one embodiment, it is i-1 according to the numbering, the second gray value of the pixel of j, numbering are i+1, j
The second gray value of pixel, numbering be i, the second gray value of the pixel of j-1 and numbering are i, the of the pixel of j+1
Two gray values, it is i that solution obtains the numbering, the horizontal direction gradient component and Vertical Square of the second gray value of the pixel of j
To gradient component, including：
It is i using gradient operator to the numbering, the level side of the second gray value of the pixel of j in the indicator card
Convolution algorithm is done to vertical direction, it is i to obtain the numbering, the horizontal direction gradient point of the second gray value of the pixel of j
Amount I (i+1, j)-I (i-1, j) and the numbering is i, vertical gradient component I (i, the j of the second gray value of the pixel of j
+ 1)-I (i, j-1), wherein, (i+1, is j) i+1 for the numbering to I, the second gray value of the pixel of j, and (i-1 is j) institute to I
It is i-1 to state numbering, the second gray value of the pixel of j, and I (i, j+1) is i for the numbering, the second gray scale of the pixel of j+1
Value, I (i, j-1) are i for the numbering, the second gray value of the pixel of j-1.
In one embodiment, it is i according to the numbering, the horizontal direction gradient of the second gray value of the pixel of j
Component and vertical gradient component, it is i, the gradient magnitude of the second gray value of the pixel of j and gradient that solution obtains numbering
The angle in direction, including：
It is i, the horizontal direction gradient component of the second gray value of the pixel of j and vertical gradient according to the numbering
Component, using arctan function, by below equation, it is i that solution obtains the numbering, the gradient magnitude and ladder of the pixel of j
The angle in degree direction：
θ (i, j)=arctan ((I (i+1, j)-I (i-1, j))/(I (i, j+1)-I (i, j-1)))
Wherein, R (i, j) is i for the numbering, the gradient magnitude of the pixel of j, θ (i, j) for it is described be i, j for numbering
Pixel gradient direction angle, I (i+1, is j) i+1 for the numbering, the second gray value of the pixel of j, I (i-1,
J) it is i-1 for the numbering, the second gray value of the pixel of j, I (i, j+1) they are i for the numbering, the of the pixel of j+1
Two gray values, I (i, j-1) are i for the numbering, the second gray value of the pixel of j-1.
In one embodiment, according to the angle of the gradient magnitude and gradient direction of each pixel is obtained
The gradient orientation histogram of indicator card, including：
The indicator card is divided into into multiple grids；
Statistics obtains the gradient orientation histogram of each grid in the plurality of grid；
Described in series connection, the gradient orientation histogram of each grid, obtains the gradient orientation histogram of the indicator card.
In one embodiment, statistics obtains the gradient direction of current grid in the plurality of grid in such a way
Histogram：
According to the angle of the gradient direction of each pixel in the current grid, by the pixel in the current grid
Point, is partitioned in multiple different angular ranges；
Statistics obtains the gradient magnitude of all angles scope in the plurality of different angular ranges, wherein, described each angle
The gradient magnitude of degree scope is obtained in such a way：The gradient magnitude of each pixel in the range of current angular is entered
Row is cumulative, using accumulation result as the current angular scope gradient magnitude；
According to the gradient magnitude of all angles scope in the plurality of angular range, the gradient side of the current grid is obtained
To histogram.
In one embodiment, by the pixel in the current grid, it is partitioned in multiple different angular ranges,
Including：
According to the angle of the gradient direction of each pixel in the indicator card, by the pixel in the current grid,
It is partitioned in 4 different angular ranges.
In one embodiment, the gradient orientation histogram according to the indicator card, recognizes the indicator card
Type, including：
Using the gradient orientation histogram of the indicator card as the indicator card identification feature；
It is input into the identification feature as the input data of default grader into the default grader, obtains described
The type of indicator card.
In one embodiment, the default grader is obtained in such a way：
The indicator card of multiple identified types is obtained as training sample；
Training sample input preliminary classification device is trained, the default grader is obtained.
Based on identical thinking, the embodiment of the present invention additionally provides a kind of identifying device of indicator card, including：
First acquisition module, for the first gray value according to each pixel in indicator card to be identified, shows work(to described
Figure carries out binary conversion treatment, obtains the second gray value of each pixel in the indicator card；
Second acquisition module, for the second gray value according to each pixel, obtains each pixel
The angle of gradient magnitude and gradient direction；
Module is solved, for the angle of gradient magnitude and gradient direction according to each pixel, solution obtains institute
State the gradient orientation histogram of indicator card；
Identification module, for the gradient orientation histogram according to the indicator card, recognizes the type of the indicator card.
In embodiments of the present invention, the gradient orientation histogram of indicator card is obtained by indicator card figure itself, is recycled
The gradient orientation histogram identification indicator card type of indicator card, efficiently solves recognition efficiency present in existing recognition methods low
The technical problem low with recognition accuracy.
Description of the drawings
Accompanying drawing described herein is used for providing a further understanding of the present invention, constitutes the part of the application, not
Constitute limitation of the invention.In the accompanying drawings：
Fig. 1 is the process chart of indicator card recognition methods according to embodiments of the present invention；
Fig. 2 is the gradient orientation histogram of the acquisition indicator card in indicator card recognition methods according to embodiments of the present invention
Process chart；
Fig. 3 is the structural representation of indicator card identifying device according to embodiments of the present invention.
Specific embodiment
It is for making the object, technical solutions and advantages of the present invention become more apparent, with reference to embodiment and accompanying drawing, right
The present invention is described in further details.Here, the exemplary embodiment of the present invention and its illustrate for explaining the present invention, but and
It is not as a limitation of the invention.
In view of existing indicator card recognition methods, during feature extraction, due to the limitation of method itself, one
As can only extract some primary characteristic quantities, and the result for obtaining is identified to indicator card using these primary characteristic quantities
The degree of accuracy it is often not high.Meanwhile, and as existing method needs to be processed according to the former data of indicator card mostly, and it is different
The method that obtains of indicator card and process preserving type and differ, there is very big difference in the process to different indicator cards original data accordingly
It is different, often so recognition result can be made using the former data that mode processes each indicator card are uniformly processed during concrete process
Into error.Additionally, existing recognition methods, such as vector characteristic method and moment characteristics method etc., due to the problem of method itself design,
When being embodied as, it is vulnerable to the impact of noise, causes recognition result accuracy.For producing the basic original of these problems
Cause, it is considered to can be identified to indicator card according to indicator card figure itself, such that it is able to avoid using former data.Further,
Consider specifically to recognize the type of indicator card using the gradient orientation histogram of indicator card, so as to reach high efficiency, Gao Zhun
The purpose of exactness ground identification indicator card.
Based on above-mentioned thinking thinking, a kind of indicator card recognition methods is embodiments provided.Refer to Fig. 1, the party
Method may comprise steps of：
Step 101：According to the first gray value of each pixel in indicator card to be identified, indicator card is carried out at binaryzation
Reason, obtains the second gray value of each pixel in indicator card.
In above-mentioned embodiment, the indicator card can be the envelope being made up of with the relation curve that displacement is gradually changed load
Closed curve.For representing a suction period internal load of oil pumper and the change of displacement.Wherein, surface dynamometer card is also referred to as light
Bar indicator card, is the closed curve for representing polished rod load with suspension point change in displacement, and abscissa is suspension point displacement, and ordinate is suspension point
Load.The working condition of oil pumper understands direct reaction on the pump dynagraoph of down-hole, while also can be reflected on surface dynamometer card indirectly.
Additionally, for the dead load born of rod string is only considered, the indicator card drawn out is the parallelogram of rule, can be with
Referred to as theoretical indicator card.
In above-mentioned embodiment, gray value may generally refer to the color depth at black white image midpoint, scope typically from 0 to
255, white is 255, and black is 0, therefore black and white picture is also referred to as gray level image.Specifically, gray value is in medical science, field of image recognition
Have been widely used.
In an embodiment of the application, due to difference between the first gray value of each pixel in indicator card image
Than larger, therefore can not directly using the first gray value of each pixel, so needing the first gray scale of each pixel
Value carries out process and obtains corresponding second gray value, is subsequently located such that it is able to the second gray value using each pixel
Reason.The first gray value described in above-mentioned embodiment according to each pixel in indicator card, carries out two-value to indicator card image
Change is processed, and is obtained the second gray value of each pixel in indicator card, is embodied as including：：
S1：Second gray value of first gray value in described each pixel more than or equal to the pixel of predetermined threshold value is set
For 1；
S2：Second gray value of first gray value in described each pixel less than the pixel of the predetermined threshold value is set
For 0.
It should be noted that the predetermined threshold value in above-mentioned embodiment can be that each pixel first is grey in the indicator card
The geometrical mean of the arithmetic mean of instantaneous value of angle value, or each the first gray value of pixel, can also be each pixel
The parameters such as the weighted average of the first gray value.For above-mentioned predetermined threshold value, can flexibly choose according to situation is embodied as, it is right
This, the application is not construed as limiting.
When being embodied as, above-mentioned embodiment can also carry out binaryzation according to formula given below to each pixel
Process：
Wherein, T is predetermined threshold value,It is i for numbering, the first gray value of the pixel of j, Ii,jIt is i for numbering, the picture of j
Second gray value of vegetarian refreshments, i are that 1 to the integer between n, j is that 1 to the integer between m, n × m is pixel in the indicator card
Number.
Step 102：According to the second gray value of each pixel, solution obtains gradient magnitude and the gradient of each pixel
The angle in direction.
In above-mentioned embodiment, described gradient can be one vectorial, including gradient magnitude and gradient direction.In scalar
In, the gradient on certain point points to the fastest-rising direction of scalar field.Wherein, the length of gradient is the change of this maximum
Rate.Strictly speaking, from Euclidean space RnGradient to the function of R is in RnThe optimal linear approximation in certain point.In this meaning
In justice, gradient can also be special circumstances of Jacobian matrix.
In an embodiment of the application, in order to solve the gradient orientation histogram for obtaining indicator card, need first
Obtain the gradient of each pixel in the indicator card.As general gradient is mostly a vector, i.e. it is right that gradient includes
The gradient magnitude answered and the angle of corresponding gradient direction.Therefore, it is in order to obtain the gradient of each pixel, corresponding to need really
The angle of the gradient magnitude and gradient direction of fixed each pixel.Specific enforcement is referred to following steps and asks for each pixel
In point, numbering is i, the gradient magnitude of the second gray value of the pixel of j and the angle of gradient direction：
S1：It is i-1 according to numbering, the second gray value of the pixel of j, numbering are i+1, the second gray scale of the pixel of j
Value, numbering are i, and the second gray value of the pixel of j-1 and numbering are i, and the second gray value of the pixel of j+1, solution are obtained
Numbering is i, the horizontal direction gradient component and vertical gradient component of the second gray value of the pixel of j；
S2：It is i, the horizontal direction gradient component of the second gray value of the pixel of j and vertical gradient according to numbering
Component, it is i, the gradient magnitude of the second gray value of the pixel of j and the angle of gradient direction that solution obtains numbering；
Wherein, i be 1 to the integer between n, j be 1 to the integer between m, n × m be in the indicator card pixel
Number.
In an embodiment of the application, in order to the second gray value according to each pixel obtains each pixel
The second gray value horizontal direction gradient component and vertical gradient component, can perform in such a way：Using ladder
Degree operator is i to numbering, and the second gray value of the pixel of j horizontally and vertically does convolution fortune in indicator card
Calculate, it is i to obtain numbering, the horizontal direction gradient component I of the second gray value of the pixel of j (i+1, j)-I (i-1, j) and numbering
For i, vertical gradient component I (i, j+1)-I (i, j-1) of the second gray value of the pixel of j, wherein, (i+1 j) is I
Numbering is i+1, the second gray value of the pixel of j, and (i-1, is j) i-1 for numbering to I, the second gray value of the pixel of j, I
(i, j+1) is i for numbering, the second gray value of the pixel of j+1, and I (i, j-1) is i for numbering, the second of the pixel of j-1
Gray value.
In an embodiment of the application, in order to preferably obtain the second gray value of each pixel in indicator card
Horizontal direction gradient component and vertical gradient component, it may be considered that the graphic feature of indicator card, using corresponding gradient
Operator solves the gradient magnitude of the second gray value of each pixel and the angle of gradient direction.Wherein, above-mentioned corresponding gradient
Operator is specifically as follows [- 1,0,1].Certainly, solved simply one kind with [- 1,0,1] as gradient operator to schematically illustrate,
Other can also be taken and corresponding horizontal direction gradient component and vertical gradient can be obtained according to the second gray value of pixel
The data of component are used as gradient operator, in this regard, the application is not construed as limiting.
In an embodiment of the application, in order to solve, to obtain numbering be i, the gradient magnitude and ladder of the pixel of j
The angle in degree direction, can be i according to numbering, the horizontal direction gradient component and Vertical Square of the second gray value of the pixel of j
To gradient component, it is i to solve numbering, and the angle of the gradient magnitude and gradient direction of the pixel of j, concrete which can be wrapped
Include：
It is i, the horizontal direction gradient component of the second gray value of the pixel of j and vertical gradient point according to numbering
Amount, using arctan function, by below equation, it is i that solution obtains numbering, the gradient magnitude and gradient direction of the pixel of j
Angle：
θ(xi,yi)=arctan ((I (xi+1,yi)-I(xi-1,yi))/(I(xi,yi+1)-I(xi,yi-1))),
Wherein, R (i, j) is i for numbering, and the gradient magnitude of the pixel of j, θ (i, j) are i for numbering, the pixel of j
The angle of gradient direction, I (i+1, is j) i+1 for numbering, the second gray value of the pixel of j, I (i-1, is j) i-1 for numbering,
Second gray value of the pixel of j, I (i, j+1) are i for numbering, and the second gray value of the pixel of j+1, I (i, j-1) are volume
Number be i, the second gray value of the pixel of j-1.
In a specific embodiment, binary conversion treatment is carried out to the indicator card of a pumpingh well, is obtained in the indicator card
After second gray value of each pixel, according to the second gray value of each pixel, using what is provided in above-mentioned embodiment
Computing formula, solution obtain the gradient magnitude of the second gray value of each pixel in the indicator card and the angle of gradient direction.
Specifically, as the indicator card after binaryzation is respectively along X-direction and along the gradient span of Y direction：{ -1,0,1 },
Therefore correspondingly, in the indicator card, the angular configurations scope of the gradient direction of each pixel is then：0 °, and ± 45 °, ± 90 °,
± 135 °, 180 ° }, meanwhile, the span of gradient magnitude is：{0,1,1.41}.
Step 103：According to the angle of the gradient magnitude and gradient direction of each pixel, solution obtains the gradient of indicator card
Direction histogram.
In an embodiment of the application, in order to solve the gradient orientation histogram for obtaining indicator card, generally require
The indicator card is divided into into multiple grids first, the histogram of each grid is solved respectively, further according to the histogram of each grid, finally
Obtain the gradient orientation histogram of the indicator card.Specifically, Fig. 2 can be referred to.According to steps of processing：
Step 201：The indicator card is divided into into multiple grids；
Step 202：By the pixel in each grid in multiple grids, according to the angle of the gradient direction of pixel,
It is partitioned into multiple angular ranges；
Step 203：The gradient magnitude of each pixel in all angles scope in each grid is counted respectively, obtains each
The gradient magnitude of all angles scope in grid；
Step 204：According to the gradient magnitude of all angles scope in each grid, the gradient direction for obtaining each grid is straight
Fang Tu；
Step 205：Connect the gradient orientation histogram of each grid, obtain the gradient orientation histogram of the indicator card.
It should be noted that counting in each grid each picture in all angles scope described in above-mentioned embodiment respectively
The gradient magnitude of vegetarian refreshments, can add the gradient magnitude of the pixel in each angular range in each grid respectively, will be tired
Plus result is used as the gradient magnitude of the angular range.
Gradient magnitude described in above-mentioned embodiment according to all angles scope in each grid, obtains each grid
Histogram of gradients, can be will in each grid the gradient magnitude of all angles scope arrange unified to one it is vectorial in, i.e.,
For having obtained the histogram of gradients of the grid.
Connect in above-mentioned embodiment the gradient orientation histogram of each grid, obtain the gradient direction Nogata of the indicator card
Figure, can be by the gradient orientation histogram of each grid in the indicator card be connected in series to one it is vectorial in, as obtained this and shown
The gradient orientation histogram of work(figure.
Above-mentioned embodiment can also be specifically：The indicator card is divided into into multiple grids；
According to the multiple grids of below step process, the pixel in current grid is drawn according to the angle in the direction of its gradient
It is divided into multiple angular ranges；
Count the gradient magnitude of all angles scope in multiple angular ranges, wherein, current angular in all angles scope
The gradient magnitude of scope is obtained again in such a way：The gradient magnitude of each pixel in current angular scope is carried out
It is cumulative, using accumulation result as current angular scope gradient magnitude；
According to the gradient magnitude of all angles scope in multiple angular ranges, the gradient direction Nogata of current grid is generated
Figure；
Connect the gradient orientation histogram of each grid, obtain the gradient orientation histogram of the indicator card.
In a specific embodiment, the indicator card of an oil pumper is divided into 4 grids, by the pixel in each grid
Point is further divided into n angular range according to the angle of their gradient direction.All angles in each grid are counted respectively
The gradient magnitude of each pixel in scope, the gradient magnitude of all pixels point in same angular range is added up, is worked as
The gradient magnitude of front angular range.Further according to the gradient magnitude of all angles scope, the gradient direction Nogata of correspondence grid is obtained
Figure, and then obtain the gradient orientation histogram of the indicator card.For example, for the first grid, the gradient magnitude of first angle scope
For A1, the gradient magnitude of second angle scope is A2, and the gradient magnitude of the n-th angular range is An.In the same manner, it is right
In the second grid, the gradient magnitude of all angles scope is respectively：B1, B2, Bn.For third party's lattice, respectively
The gradient magnitude of individual angular range is respectively：C1, C2, Cn.For square grids, the ladder of all angles scope
Degree amplitude is respectively：D1, D2, Dn.Further according to the gradient magnitude of all angles scope in each grid, obtain
The gradient orientation histogram of each grid.For example, for the first grid, the gradient orientation histogram of the grid for (A1,
A2, An).For the second grid, the gradient orientation histogram of the grid for (B1, B2,
Bn).For third party's lattice, the gradient orientation histogram of the grid is (C1, C2, Cn).For square grids,
The gradient orientation histogram of the grid is (D1, D2, Dn).Finally, connect each grid gradient direction it is straight
Fang Tu, obtains the gradient orientation histogram of the indicator card.Specifically, the gradient orientation histogram of above-mentioned 4 grids is connected on
During one is vectorial, obtain vector (A1, A2, An, B1, B2, Bn, C1,
C2, Cn, D1, D2, Dn) it is the gradient orientation histogram of the indicator card.
In an embodiment of the application, it is contemplated that the graphics feature of handled pumping-unit workdone graphic, can be by
The all pixels in each grid in multiple grids press the angle of the gradient direction of photograph vegetarian refreshments, are partitioned into multiple angle models
When enclosing, typically by all pixels point of each grid according to the angle of gradient direction, 4 can be divided into as the case may be
Individual angular range, to ensure the accuracy of subsequent treatment.Certainly, division methods recited herein are intended merely to more clearly say
The bright embodiment of the present invention, can also select the suitable division methods that meet the requirements when being embodied as the case may be, in this regard,
The application is not construed as limiting.
In a specific embodiment, taken according to the angle of the gradient direction of each pixel in a pumping-unit workdone graphic
Value scope { 0 °, ± 45 °, ± 90 °, ± 135 °, 180 ° }, by the pixel in each grid according to gradient direction angular range
4 angular ranges are divided into, respectively：0 ° and 180 ° of first angular range correspondence ,+45 ° of second angular range correspondence with-
135 °, the 3rd angular range is corresponding ± 90 °, and the 4th angular range is corresponding+135 ° and -45 °.
Step 104：According to the gradient orientation histogram of indicator card, the type of the indicator card is recognized.
In an embodiment of the application, in order to work(is shown in the gradient orientation histogram identification using the indicator card for obtaining
The type of figure, specifically can be processed in such a way：
S1：Using the gradient orientation histogram of indicator card as the indicator card feature；
S2：According to indicator card feature, the type of the indicator card is recognized using grader.
It should be noted that the grader in above-mentioned embodiment is to learn a classification letter on the basis of data with existing
Number constructs a disaggregated model, and this disaggregated model is described grader (Classifier).Wherein, the function or
Model the data recording in database can be mapped in given classification some, such that it is able to be applied to data prediction.
In a word, grader is the general designation of the method classified to sample in data mining, comprising decision tree, logistic regression, simple shellfish
Ye Si, neutral net scheduling algorithm.What the grader involved by the embodiment of the present invention was mainly applied is neural network algorithm.Wherein,
Neutral net (Artificial Neural Networks, ANN) algorithm, is the second way for simulating people's thinking.This is one
Individual Kind of Nonlinear Dynamical System, its characteristic are that the distributed storage of information and concurrent collaborative are processed.Although single neuron
Structure is extremely simple, and function is limited, but the behavior achieved by the network system of a large amount of neuron compositions is extremely abundant multiple
Miscellaneous.The embodiment of the present invention grader based on nerual network technique specifically related to can be SVM (Support Vector
Machine, SVMs) grader, or RBF (Radial basis function, RBF) grader
Deng.However, it is necessary to explanation, above-mentioned cited grader is in order to present embodiment is better described, when being embodied as,
The grader suitably based on nerual network technique can be selected to carry out type identification, in this regard, the application is not as the case may be
It is construed as limiting.
In an embodiment of the application, before the type of indicator card is recognized using grader, the method is also wrapped
Include：By being input into the indicator card of multiple identified types to grader, to be trained to grader.
In the present embodiment, as the grader for being used is the grader based on neural network algorithm, such classification
Utensil has certain learning and memory ability.Before type identification is carried out using such grader, generally require by input
Substantial amounts of sample is trained, learns so that grader can set up an accurate classification of type model, such that it is able to root
According to the gradient orientation histogram of the indicator card of input, by the classification of type model having built up, the type of the indicator card is recognized.
In a specific embodiment, type knowledge is being carried out to the indicator card of oil pumper to be identified using SVM classifier
Before not, the indicator card input SVM classifier of the oil pumper of a large amount of marked good types is trained.SVM classifier according to
The data of the indicator card of these marked good types set up the classification of type model of pumping-unit workdone graphic.It is to be identified receiving
After the gradient orientation histogram of pumping-unit workdone graphic, the gradient orientation histogram correspondence can be judged by classification of type model
Pumping-unit workdone graphic which type belonged to, so as to complete the identification to the indicator card.
In embodiments of the present invention, by the gradient orientation histogram for obtaining indicator card thus according to indicator card figure itself,
The gradient orientation histogram identification indicator card type of indicator card is recycled, therefore solves identification present in existing recognition methods
Efficiency is low and the low technical problem of recognition accuracy.
Based on same inventive concept, a kind of indicator card identifying device in the embodiment of the present invention, is additionally provided, such as following reality
Apply described in example.Due to the principle of indicator card identifying device solve problem it is similar to indicator card recognition methods, therefore indicator card identification
The enforcement of device may refer to the enforcement of indicator card recognition methods, repeats part and repeats no more.Used below, term is " single
Unit " or " module " can realize the combination of the software and/or hardware of predetermined function.Although the device described by following examples
Preferably with software realizing, but hardware, or the realization of the combination of software and hardware is also what is may and be contemplated.Fig. 3
It is the identification of embodiment of the present invention indicator card. a kind of structured flowchart of device, refering to shown in Fig. 3, the device can include：First obtains
Delivery block 301, the second acquisition module 302, solution module 303 and identification module 304.Below the structure is illustrated.
First acquisition module 301, for the first gray value according to each pixel in indicator card, enters to indicator card image
Row binary conversion treatment, obtains the second gray value of each pixel in indicator card；
Second acquisition module 302, for the second gray value according to each pixel, obtains the gradient width of each pixel
The angle of value and gradient direction；
Module 303 is solved, for the angle of gradient magnitude and gradient direction according to each pixel, solution obtains showing work(
The gradient orientation histogram of figure；
Identification module 304, for the gradient orientation histogram according to indicator card, recognizes the type of indicator card.
In one embodiment, obtain corresponding to solve according to the first gray value of each pixel in indicator card
Second gray value, the first acquisition module 301 can carry out binary conversion treatment according to following steps to indicator card, obtain in indicator card
Second gray value of each pixel：
S1：Second gray value of first gray value in described each pixel more than or equal to the pixel of predetermined threshold value is set
For 1；
S2：Second gray value of first gray value in described each pixel less than the pixel of the predetermined threshold value is set
For 0.
In one embodiment, for the second gray value according to each pixel in indicator card, solution obtains corresponding
The second gray value gradient magnitude and the angle of gradient direction, the second acquisition module 302 can be according to steps of processing：
S1：It is i-1 according to numbering, the second gray value of the pixel of j, numbering are i+1, the second gray scale of the pixel of j
Value, numbering are i, and the second gray value of the pixel of j-1 and numbering are i, and the second gray value of the pixel of j+1, solution are obtained
Numbering is i, the horizontal direction gradient component and vertical gradient component of the second gray value of the pixel of j；
S2：It is i, the horizontal direction gradient component of the second gray value of the pixel of j and vertical gradient according to numbering
Component, it is i, the gradient magnitude of the second gray value of the pixel of j and the angle of gradient direction to solve numbering.
In above-mentioned embodiment, in order to solve, to obtain in each pixel numbering be i to the second acquisition module 302, the pixel of j
The horizontal direction gradient component and vertical gradient component of the second gray value of point, it is concrete to locate in such a way again
Reason：Be i using gradient operator to numbering, the second gray value of the pixel of j in indicator card horizontally and vertically
Do convolution algorithm, it is i to obtain numbering, the horizontal direction gradient component I of the second gray value of the pixel of j (i+1, j)-I (i-1,
J) and numbering is i, vertical gradient component I (i, j+1)-I (i, j-1) of the second gray value of the pixel of j, wherein, I (i
+ 1, it is j) i+1 for numbering, the second gray value of the pixel of j, (i-1, is j) i-1 for numbering to I, and the second of the pixel of j is grey
Angle value, I (i, j+1) are i for numbering, the second gray value of the pixel of j+1, and I (i, j-1) is i for numbering, the pixel of j-1
The second gray value.
In above-mentioned embodiment, the second acquisition module 302 obtains the gradient of each the second gray value of pixel to solve
Amplitude and gradient direction angle, specifically can be performed in such a way：It is i according to numbering, the second gray value of the pixel of j
Horizontal direction gradient component and vertical gradient component, using arctan function, by below equation, solution is numbered
For i, the angle of the gradient magnitude and gradient direction of the pixel of j：
θ(xi,yi)=arctan ((I (xi+1,yi)-I(xi-1,yi))/(I(xi,yi+1)-I(xi,yi-1)))
Wherein, R (i, j) is i for numbering, and the gradient magnitude of the pixel of j, θ (i, j) are i for numbering, the pixel of j
The angle of gradient direction, I (i+1, is j) i+1 for numbering, the second gray value of the pixel of j, I (i-1, is j) i-1 for numbering,
Second gray value of the pixel of j, I (i, j+1) are i for numbering, and the second gray value of the pixel of j+1, I (i, j-1) are volume
Number be i, the second gray value of the pixel of j-1.
In one embodiment, module 303 is solved in order to solve the gradient orientation histogram for obtaining indicator card, specifically may be used
With according to steps of processing：
S1：Indicator card is divided into into multiple grids；
S2：By the pixel in each grid in multiple grids according to its gradient direction angular divisions be multiple angles
Scope；
S3：The gradient magnitude of each pixel in all angles scope in each grid is counted respectively, obtains each grid
The gradient magnitude of middle all angles scope；
S4：According to the gradient magnitude of all angles scope in each grid, the gradient orientation histogram of each grid is obtained；
S5：Connect the gradient orientation histogram of each grid, obtain the gradient orientation histogram of indicator card.
In above-mentioned embodiment, it is contemplated that the graphic feature of indicator card, specifically can be according to each pixel in indicator card
Gradient direction angular range, the pixel in each grid in multiple grids is divided into into 4 angular ranges.
In one embodiment, identification module 304 recognizes indicator card for the gradient orientation histogram according to indicator card
Type, specifically can perform in such a way：
S1：Using the gradient orientation histogram of indicator card as indicator card feature；
S2：According to indicator card feature, the type of indicator card is recognized using grader.
In above-mentioned embodiment, in order to elder generation sets up indicator card classification of type model in Classification and Identification device, need to instruct in advance
Practice grader.That is, before the type of indicator card is recognized using grader, first pass through to the grader and be input into multiple identified classes
The indicator card of type, is trained to the grader, to set up indicator card classification of type model.Wherein, above-mentioned grader is to answer
With the grader of nerual network technique.Can be specifically, SVM classifier, or RBF graders etc..However, it is necessary to illustrate
, above-mentioned cited grader is in order to present embodiment is better described, when being embodied as, can be according to concrete feelings
Condition, selects the grader suitably based on nerual network technique to carry out type identification, in this regard, the application is not construed as limiting.
In one specific embodiment, using the indicator card recognition methods/device of the application offer to 1138 indicator cards
It is identified, indicator card recognition methods and device raising while recognition accuracy is improved provided by the application is provided
The speed of identification.
In terms of feature extraction, it is Intel (R) Core (TM) i3-2130CPU, dominant frequency 3.40GHz, interior saves as 4G's in CPU
On the machine of configuration, simple gradient direction histogram and one are extracted respectively to indicator card image of 1138 resolution ratio for 96x128
As gradient orientation histogram feature, and count and extract the time-consuming of feature, repeat 100 tests, average, as a result such as table 1：
Table 1
From experimental result as can be seen that the computational efficiency of the two-value gradient orientation histogram of the present invention is general gradient direction
It is histogrammic more than 5 times.
In terms of indicator card recognition accuracy, in the test set for including 10 kinds of different types, have 938 indicator card samples altogether
On, the recognition accuracy such as table 2 below of the present invention：
Table 2
From the graph experimental result can be seen that the two-value gradient orientation histogram feature of the present invention indicator card identification it is accurate
Rate is substantially better than other features.
As can be seen from the above description, the embodiment of the present invention realizes following technique effect：Due to utilizing indicator card
Gradient orientation histogram recognize the type of indicator card, therefore it is low to there is recognition efficiency in solving existing indicator card recognition methods
The technical problem low with recognition accuracy, realizes high efficiency, accurately recognizes；Due to can be according to indicator card figure itself
Obtain the gradient orientation histogram of the indicator card for type identification, it is to avoid using former data, reduce the difference of former data
To the error caused by recognition result；Additionally, due also to utilizing the grader based on nerual network technique according to the ladder of indicator card
Degree histogram recognizes the type of indicator card, further increases the degree of accuracy of indicator card identification, while improve the efficiency of identification.
Obviously, those skilled in the art should be understood that each module or each step of the above-mentioned embodiment of the present invention can be with
Realized with general computing device, they can be concentrated on single computing device, or be distributed in multiple computing devices
On the network for being constituted, alternatively, they can be realized with the executable program code of computing device, it is thus possible to by it
Store in the storage device by computing device performing, and in some cases, can be holding different from order herein
The shown or described step of row, or they are fabricated to each integrated circuit modules respectively, or will be multiple in them
Module or step are fabricated to single integrated circuit module to realize.So, the embodiment of the present invention is not restricted to any specific hard
Part and software are combined.
The preferred embodiments of the present invention are the foregoing is only, the present invention is not limited to, for the skill of this area
For art personnel, the embodiment of the present invention can have various modifications and variations.It is all within the spirit and principles in the present invention, made
Any modification, equivalent substitution and improvements etc., should be included within the scope of the present invention.
Claims (11)
1. a kind of indicator card recognition methods, it is characterised in that include：
According to the first gray value of each pixel in indicator card to be identified, binary conversion treatment is carried out to the indicator card, is obtained
Second gray value of each pixel in the indicator card；
According to the second gray value of each pixel, solution obtains the gradient width of the second gray value of each pixel
The angle of value and gradient direction；
According to the angle of the gradient magnitude and gradient direction of each pixel, solution obtains the gradient direction of the indicator card
Histogram；
According to the gradient orientation histogram of the indicator card, the type of the indicator card is recognized.
2. method according to claim 1, it is characterised in that described according to of each pixel in indicator card to be identified
One gray value, carries out binary conversion treatment to the indicator card, obtains the second gray value of each pixel in the indicator card, bag
Include：
Second gray value of first gray value in described each pixel more than or equal to the pixel of predetermined threshold value is set to into 1；
Second gray value of first gray value in described each pixel less than the pixel of the predetermined threshold value is set to into 0.
3. method according to claim 2, it is characterised in that according to the second gray value of each pixel, solves
Gradient magnitude and the angle of gradient direction of the second gray value of each pixel is obtained, including：
In such a way, it is i to number in asking for described each pixel, the gradient magnitude of the second gray value of the pixel of j
With the angle of gradient direction：
It is i-1 according to numbering, the second gray value of the pixel of j, numbering are i+1, the second gray value of the pixel of j, numbering
For i, the second gray value of the pixel of j-1 and numbering are i, and the second gray value of the pixel of j+1, solution obtain the numbering
For i, the horizontal direction gradient component and vertical gradient component of the second gray value of the pixel of j；
It is i, the horizontal direction gradient component of the second gray value of the pixel of j and vertical gradient point according to the numbering
Amount, it is i, the gradient magnitude of the second gray value of the pixel of j and the angle of gradient direction that solution obtains numbering；
Wherein, i is that 1 to the integer between n, j is that 1 to the integer between m, n × m is the number of pixel in the indicator card.
4. method according to claim 3, it is characterised in that be i-1 according to the numbering, the second of the pixel of j is grey
Angle value, numbering are i+1, and the second gray value of the pixel of j, numbering are i, and the second gray value of the pixel of j-1 and numbering are
Second gray value of the pixel of i, j+1, it is i that solution obtains the numbering, the horizontal direction of the second gray value of the pixel of j
Gradient component and vertical gradient component, including：
Be i using gradient operator to the numbering, the horizontal direction of the second gray value of the pixel of j in the indicator card and
Vertical direction does convolution algorithm, and it is i, the horizontal direction gradient component I (i of the second gray value of the pixel of j to obtain the numbering
+ 1, j)-I (i-1, j) and it is described numbering be i, vertical gradient component I (i, the j+1)-I of the second gray value of the pixel of j
(i, j-1), wherein, (i+1, is j) i+1 for the numbering to I, the second gray value of the pixel of j, and (i-1 j) is the numbering to I
For i-1, the second gray value of the pixel of j, I (i, j+1) are i for the numbering, the second gray value of the pixel of j+1, I
(i, j-1) is i for the numbering, the second gray value of the pixel of j-1.
5. method according to claim 4, it is characterised in that be i according to the numbering, the second gray scale of the pixel of j
The horizontal direction gradient component and vertical gradient component of value, it is i that solution obtains numbering, the second gray value of the pixel of j
Gradient magnitude and gradient direction angle, including：
It is i, the horizontal direction gradient component of the second gray value of the pixel of j and vertical gradient point according to the numbering
Amount, using arctan function, by below equation, it is i, the gradient magnitude of the pixel of j and gradient that solution obtains the numbering
The angle in direction：
θ (i, j)=arctan ((I (i+1, j)-I (i-1, j))/(I (i, j+1)-I (i, j-1)))
Wherein, R (i, j) is i for the numbering, the gradient magnitude of the pixel of j, θ (i, j) for it is described be i for numbering, the picture of j
The angle of the gradient direction of vegetarian refreshments, (i+1, is j) i+1 for the numbering to I, the second gray value of the pixel of j, and (i-1 j) is I
The numbering is i-1, the second gray value of the pixel of j, and I (i, j+1) is i for the numbering, and the second of the pixel of j+1 is grey
Angle value, I (i, j-1) are i for the numbering, the second gray value of the pixel of j-1.
6. method according to claim 1, it is characterised in that according to gradient magnitude and the gradient side of each pixel
To angle obtain the gradient orientation histogram of the indicator card, including：
The indicator card is divided into into multiple grids；
Statistics obtains the gradient orientation histogram of each grid in the plurality of grid；
Described in series connection, the gradient orientation histogram of each grid, obtains the gradient orientation histogram of the indicator card.
7. method according to claim 6, it is characterised in that in such a way statistics obtain in the plurality of grid when
The gradient orientation histogram of front lattice：
According to the angle of the gradient direction of each pixel in the current grid, by the pixel in the current grid, draw
Divide to enter in multiple different angular ranges；
Statistics obtains the gradient magnitude of all angles scope in the plurality of different angular ranges, wherein, all angles model
The gradient magnitude for enclosing is obtained in such a way：The gradient magnitude of each pixel in the range of current angular is tired out
Plus, using accumulation result as the current angular scope gradient magnitude；
According to the gradient magnitude of all angles scope in the plurality of angular range, the gradient direction for obtaining the current grid is straight
Fang Tu.
8. method according to claim 7, it is characterised in that by the pixel in the current grid, be partitioned into multiple
In different angular ranges, including：
According to the angle of the gradient direction of each pixel in the indicator card, by the pixel in the current grid, divide
Enter in 4 different angular ranges.
9. method according to claim 1, it is characterised in that the gradient orientation histogram according to the indicator card,
The type of the indicator card is recognized, including：
Using the gradient orientation histogram of the indicator card as the indicator card identification feature；
It is input into the identification feature as the input data of default grader into the default grader, obtains described showing work(
The type of figure.
10. method according to claim 9, it is characterised in that the default grader is obtained in such a way：
The indicator card of multiple identified types is obtained as training sample；
Training sample input preliminary classification device is trained, the default grader is obtained.
11. a kind of indicator card identifying devices, it is characterised in that include：
First acquisition module, for the first gray value according to each pixel in indicator card to be identified, enters to the indicator card
Row binary conversion treatment, obtains the second gray value of each pixel in the indicator card；
Second acquisition module, for the second gray value according to each pixel, obtains the gradient of each pixel
The angle of amplitude and gradient direction；
Module is solved, for the angle of gradient magnitude and gradient direction according to each pixel, solution obtains described showing
The gradient orientation histogram of work(figure；
Identification module, for the gradient orientation histogram according to the indicator card, recognizes the type of the indicator card.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610873368.8A CN106529542A (en) | 2016-09-30 | 2016-09-30 | Indicator diagram identification method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610873368.8A CN106529542A (en) | 2016-09-30 | 2016-09-30 | Indicator diagram identification method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106529542A true CN106529542A (en) | 2017-03-22 |
Family
ID=58331389
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610873368.8A Pending CN106529542A (en) | 2016-09-30 | 2016-09-30 | Indicator diagram identification method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106529542A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108710920A (en) * | 2018-06-05 | 2018-10-26 | 北京中油瑞飞信息技术有限责任公司 | Indicator card recognition methods and device |
| CN108764361A (en) * | 2018-06-01 | 2018-11-06 | 北京中油瑞飞信息技术有限责任公司 | The operating mode's switch method of beam type oil pumping machine indicating diagram based on integrated study |
| CN111379555A (en) * | 2018-12-29 | 2020-07-07 | 中国石油天然气股份有限公司 | Fault diagnosis method and device for oil pumping well |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104361365A (en) * | 2014-12-04 | 2015-02-18 | 杭州和利时自动化有限公司 | Oil-well pump running condition recognition method and device |
| CN104392224A (en) * | 2014-12-04 | 2015-03-04 | 西南交通大学 | Crack detection method for road surface |
| CN104463240A (en) * | 2013-09-23 | 2015-03-25 | 深圳市朗驰欣创科技有限公司 | Method and device for controlling list interface |
| CN104866862A (en) * | 2015-04-27 | 2015-08-26 | 中南大学 | Strip steel surface area type defect identification and classification method |
| CN105426905A (en) * | 2015-11-02 | 2016-03-23 | 国网山东省电力公司电力科学研究院 | Robot barrier identification method based on gradient histogram and support vector machine |
| CN105787466A (en) * | 2016-03-18 | 2016-07-20 | 中山大学 | Vehicle type fine identification method and system |
| CN105787437A (en) * | 2016-02-03 | 2016-07-20 | 东南大学 | Vehicle brand type identification method based on cascading integrated classifier |
| CN105865811A (en) * | 2016-04-14 | 2016-08-17 | 辽宁工业大学 | Automotive suspension vibration damper energy characteristic testing system and automotive suspension vibration damper energy characteristic metering method |
| CN105930791A (en) * | 2016-04-19 | 2016-09-07 | 重庆邮电大学 | Road traffic sign identification method with multiple-camera integration based on DS evidence theory |
-
2016
- 2016-09-30 CN CN201610873368.8A patent/CN106529542A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104463240A (en) * | 2013-09-23 | 2015-03-25 | 深圳市朗驰欣创科技有限公司 | Method and device for controlling list interface |
| CN104361365A (en) * | 2014-12-04 | 2015-02-18 | 杭州和利时自动化有限公司 | Oil-well pump running condition recognition method and device |
| CN104392224A (en) * | 2014-12-04 | 2015-03-04 | 西南交通大学 | Crack detection method for road surface |
| CN104866862A (en) * | 2015-04-27 | 2015-08-26 | 中南大学 | Strip steel surface area type defect identification and classification method |
| CN105426905A (en) * | 2015-11-02 | 2016-03-23 | 国网山东省电力公司电力科学研究院 | Robot barrier identification method based on gradient histogram and support vector machine |
| CN105787437A (en) * | 2016-02-03 | 2016-07-20 | 东南大学 | Vehicle brand type identification method based on cascading integrated classifier |
| CN105787466A (en) * | 2016-03-18 | 2016-07-20 | 中山大学 | Vehicle type fine identification method and system |
| CN105865811A (en) * | 2016-04-14 | 2016-08-17 | 辽宁工业大学 | Automotive suspension vibration damper energy characteristic testing system and automotive suspension vibration damper energy characteristic metering method |
| CN105930791A (en) * | 2016-04-19 | 2016-09-07 | 重庆邮电大学 | Road traffic sign identification method with multiple-camera integration based on DS evidence theory |
Non-Patent Citations (4)
| Title |
|---|
| 王香增: "《特低渗油藏采油工艺技术》", 31 December 2013 * |
| 董志成，任强: "基于BP神经网络的减振器示功图识别方法", 《机电工程》 * |
| 赵小川 编著: "《MATLAB数字图像处理：从仿真到C/C++代码的自动生成》", 30 September 2015 * |
| 郭健: "基于特征值提取与神经网络的抽油井故障诊断", 《电子设计工程》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108764361A (en) * | 2018-06-01 | 2018-11-06 | 北京中油瑞飞信息技术有限责任公司 | The operating mode's switch method of beam type oil pumping machine indicating diagram based on integrated study |
| CN108764361B (en) * | 2018-06-01 | 2020-12-22 | 北京中油瑞飞信息技术有限责任公司 | Working condition identification method of indicator diagram of beam-pumping unit based on integrated learning |
| CN108710920A (en) * | 2018-06-05 | 2018-10-26 | 北京中油瑞飞信息技术有限责任公司 | Indicator card recognition methods and device |
| CN111379555A (en) * | 2018-12-29 | 2020-07-07 | 中国石油天然气股份有限公司 | Fault diagnosis method and device for oil pumping well |
| CN111379555B (en) * | 2018-12-29 | 2023-05-26 | 中国石油天然气股份有限公司 | Fault diagnosis method and device for oil pumping well |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110059694B (en) | Intelligent identification method for character data in complex scene of power industry | |
| Tensmeyer et al. | Analysis of convolutional neural networks for document image classification | |
| CN108446700B (en) | License plate attack generation method based on anti-attack | |
| Rahaman et al. | An efficient multilevel thresholding based satellite image segmentation approach using a new adaptive cuckoo search algorithm | |
| CN108388927A (en) | Small sample polarization SAR terrain classification method based on the twin network of depth convolution | |
| CN109005145A (en) | A kind of malice URL detection system and its method extracted based on automated characterization | |
| CN108052523A (en) | Gambling site recognition methods and system based on convolutional neural networks | |
| CN111754519B (en) | Class activation mapping-based countermeasure method | |
| CN111259397B (en) | Malware classification method based on Markov graph and deep learning | |
| CN108108753A (en) | A kind of recognition methods of check box selection state based on support vector machines and device | |
| CN112597495A (en) | Malicious code detection method, system, equipment and storage medium | |
| CN110276357A (en) | A kind of method for recognizing verification code based on convolutional neural networks | |
| CN106529542A (en) | Indicator diagram identification method and device | |
| CN110751644A (en) | Road surface crack detection method | |
| CN111738367A (en) | Part classification method based on image recognition | |
| CN112381177A (en) | Dial digital character recognition method and system based on deep learning | |
| CN105654138A (en) | Orthogonal projection and dimensionality reduction classification method and system for multidimensional data | |
| CN105654070A (en) | Low-resolution face recognition method | |
| CN117315369A (en) | Fundus disease classification method and device based on neural network | |
| CN117197543A (en) | Network anomaly detection method and device based on GMD imaging and improved ResNeXt | |
| CN116704241A (en) | Full-channel 3D convolutional neural network hyperspectral remote sensing image classification method | |
| CN112463387B (en) | Method for identifying deep learning model on local server based on GPU space-time resource consumption | |
| CN114638356A (en) | Static weight guided deep neural network back door detection method and system | |
| CN116071307A (en) | Pavement defect detection model building method, detection method, storage medium and equipment | |
| CN113657493B (en) | X-ray security inspection image contraband detection method based on attention of style characteristic channel |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170322 |
|
| RJ01 | Rejection of invention patent application after publication |