CN109359643B - Dial plate pointer identification method using computer vision - Google Patents
Dial plate pointer identification method using computer vision Download PDFInfo
- Publication number
- CN109359643B CN109359643B CN201710813475.6A CN201710813475A CN109359643B CN 109359643 B CN109359643 B CN 109359643B CN 201710813475 A CN201710813475 A CN 201710813475A CN 109359643 B CN109359643 B CN 109359643B
- Authority
- CN
- China
- Prior art keywords
- dial
- image
- area
- pointer
- straight line
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 15
- 238000001514 detection method Methods 0.000 claims abstract description 14
- 230000009466 transformation Effects 0.000 claims description 4
- 230000000877 morphologic effect Effects 0.000 claims description 2
- LEMKWEBKVMWZDU-UHFFFAOYSA-N nonanedial Chemical compound O=CCCCCCCCC=O LEMKWEBKVMWZDU-UHFFFAOYSA-N 0.000 claims 2
- 238000009432 framing Methods 0.000 claims 1
- 239000011159 matrix material Substances 0.000 description 32
- 230000006870 function Effects 0.000 description 15
- 238000012549 training Methods 0.000 description 11
- 125000004122 cyclic group Chemical group 0.000 description 6
- 238000002474 experimental method Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 3
- 230000009471 action Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000009795 derivation Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000011410 subtraction method Methods 0.000 description 2
- 238000011426 transformation method Methods 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000021615 conjugation Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000007429 general method Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000011897 real-time detection Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
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
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- 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
- G06V10/24—Aligning, centring, orientation detection or correction of the image
- G06V10/243—Aligning, centring, orientation detection or correction of the image by compensating for image skew or non-uniform image deformations
-
- 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
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a dial pointer identification method using computer vision, which comprises the steps of firstly positioning an area where a dial is located, then correcting the angle deviation of the dial, then detecting a pointer and finally reading. The invention provides that a KCF tracking algorithm is applied to the positioning of the dial area, and the detection rate is greatly improved. In addition, in the existing pointer identification algorithm, a dial is required to be horizontally placed during reading, and then the reading is obtained according to an angle method. In practice, however, the dial is not necessarily placed horizontally in the image captured by the camera. The present invention addresses this problem by using SURF to correct for angular misalignment of the dial.
Description
Technical Field
The invention belongs to the technical field of computer vision and image processing, and particularly relates to a method for tracking a jittering dial plate in real time through computer vision, correcting angle deviation caused by jitter, detecting a pointer and displaying a reading.
Background
In many plants, the meters used to detect the operational status of the equipment are still mostly pointer meters. The method for ensuring the safe operation of the plant equipment depends on the regular patrol inspection of special personnel, and emergency action is taken when the abnormal indication of the instrument is found. This makes the work of patrol personnel cumbersome and inefficient, and since the equipment is not found to be abnormal at the first time by human inspection, no action can be taken at the first time. The computer vision uses machine to replace visual organ to make measurement and judgment, and it can be used in instrument detection to raise detection efficiency and automation degree. The dial plate pictures are captured through the camera and processed by a computer at the background, the dial plate scales are detected in real time, and once the dial plate scales are abnormal, emergency measures are taken immediately. In a dial pointer identification algorithm based on computer vision, the general method is to determine the area of a dial in the whole image, then find the straight line of the pointer in the dial area, and finally read the number. When determining the dial area, the current methods include a subtraction method, a Hough circle transformation method, a rotation center positioning algorithm based on a canny operator, and the like. The subtraction method requires that two pointers are strictly aligned with dial plate images at different scales, otherwise, images are subtracted to generate a plurality of interference areas; the Hough circle transformation method and the canny operator positioning method have too large calculation amount, so that the processing is too slow, the processing time is 0.4s-0.5s per frame, the real-time processing cannot be realized, and the real-time processing is particularly important in practical application. In addition, in the existing pointer identification algorithm, the dial is required to be horizontally placed in the last reading step, so that the reading is calculated by using an angle method. However, in practical applications, the dial plate shakes along with the shaking of the machine, so that the dial plate is not horizontally placed in a picture captured by the camera.
Disclosure of Invention
In order to solve the technical problems, the invention provides a dial pointer detection method using computer vision, which can better improve the detection efficiency and the automation degree in instrument detection.
The technical scheme adopted by the invention is as follows: a method for identifying a dial indicator by using computer vision comprises the following steps:
step 1: positioning the area of the dial plate;
step 2: correcting dial angle offset;
and step 3: detecting a pointer;
and 4, step 4: and (6) reading.
The invention provides that a Kernelized Correlation Filters (KCF for short) tracking algorithm is applied to the area of a positioning table panel, so that the detection rate is greatly improved, and on average, each frame is 0.05s, and real-time detection can be realized; and SURF characteristic detection is applied to the correction of dial offset, so that dial horizontal placement is not strictly required during reading, and actual requirements can be met.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention;
FIG. 2 is a flow chart of positioning a dial region in accordance with an embodiment of the present invention;
FIG. 3 is an exemplary diagram of a dial area positioning result according to an embodiment of the present invention;
FIG. 4 is an exemplary diagram of correcting dial angle offset according to an embodiment of the present invention;
FIG. 5 is a diagram illustrating an example of pointer positioning according to an embodiment of the present invention.
Detailed Description
In order to facilitate the understanding and implementation of the present invention for those of ordinary skill in the art, the present invention is further described in detail with reference to the accompanying drawings and examples, it is to be understood that the embodiments described herein are merely illustrative and explanatory of the present invention and are not restrictive thereof.
Referring to fig. 1, the method for identifying a dial indicator using computer vision according to the present invention includes the following steps:
step 1: positioning the area of the dial plate;
due to the jitter of the machine, the video-captured meter is also jittered. In a first frame image of a video, a region where a dial is located is framed, and then the position of the dial is tracked by applying a Kernelized Correlation Filters (KCF for short).
Kernelized Correlation Filters (KCF) tracking algorithm consisting ofHenriques et al. The principles and associated definitions of the KCF tracking algorithm are described below.
The KCF algorithm utilizes a cyclic shift theory to construct a training sample of the classifier, so that training data are greatly increased, and the target position is updated in a mode of screening a candidate region with the maximum similarity to the target through a kernel function, so that efficient tracking is achieved.
(1) A cyclic matrix;
for an n × 1 vector x ═ x1,...,xn]TAnd P is an n × n permutation matrix for cyclic shift
Px=[xn,x1,x2,...,x(n-1)]TIndicating that x is cyclically shifted by 1 bit, { P }ux | u ═ 0, …, n-1} indicates that x is cyclically shifted by u bits.
The circulant matrix X may be expressed as:
x represents the 1 st row in the matrix X, i.e. the basis vector, and the 2 nd to nth rows in X are cyclically shifted from the 1 st row. From the property that the circulant matrix can be diagonalized by a Discrete Fourier Transform (DFT), it can be seen that:
where ^ denotes the discrete Fourier transform, and F is the discrete Fourier transform matrix and is a constant matrix independent of x, for any z, there isFHF=FFHI is the identity matrix, H is the conjugate transpose of the matrix, XH=(X*)T。
(2) A Gaussian kernel function;
a Radial Basis Function (RBF) is a radially symmetric scalar Function, generally defined as a monotonic Function of euclidean distance from any point x to a center x 'in space, and can be referred to as κ (x, x'). The most commonly used radial basis function is the gaussian kernel function, of the form:
wherein x' is the center of the kernel function, and delta is the width parameter of the function, and the radial action range of the function is controlled. Known nuclear correlationsThen the Gaussian kernel correlation kxx′:
(3) Performing classifier training by using a kernel function;
the process of training the classifier is mainly to pass through a training sample xiAnd its regression label yiTo find a function f (z) w that minimizes the mean square errorTz, i.e.
Where w is the parameters of the classifier, z is the image region to be detected, and λ is the regularization parameter used to control the overfitting.
Defining a kernel functionWhereinThe input to a linear problem is mapped to a function of the nonlinear feature space. Then the solution of equation (5) can be written as a linear combination for the inputs, i.e.:
wherein,
α=(K+λI)-1y (7)
where α isiFor training sample xiK is a kernel matrix of n × n and it is composed of dot products of all sample pairs, i.e.:
Kij=κ(xi,xj) (8)
then
Since the kernel matrix K is a circulant matrix (the proof process is shown in derivation 1 below), it can be derived from the properties of the circulant matrix in equation (3):
wherein k isxxIs row 1 of the circulant matrix K. Substituting the above formula (10) into the formula (7) can obtain:
and simultaneously removing conjugation on two sides to obtain:
the solution of the problem is transformed to the frequency domain based on the characteristics of the circulant matrix, thereby avoiding the process of matrix inversion. The classifier is trained to find the function f (z) ═ wTz, namely finding the optimal w, and converting the finding of w into the finding of the optimal alpha through the formula (14), so that the algorithm complexity is greatly reduced, and the operation efficiency is improved. Derivation 1: proving that the kernel matrix K is a cyclic matrix;
theorem: if K (x, x ') -K (Mx, Mx ') is K (Mx, Mx ') for any permutation matrix M, K is a circulant matrix.
And (3) proving that:
p is an n × n permutation matrix for cyclic shift:then K isij=κ(xi,xj)=κ(Pix,Pjx)=κ(P-iPix,P-iPjx)=κ(x,Pj-ix)。
Because of Pn=P0Therefore, the above formula can be written as Kij=κ(x,P(j-i)modnx), mod represents the remainder. X is satisfied due to the circulant matrix X ═ C (X) formed as in formula (2)ij=x((j-i)modn)+1That is, if the element depends only on the value of (j-i) mod n, then the matrix is a circulant matrix. Thus K is the circulant matrix. The gaussian kernel matrix satisfies the condition κ (x, x ') ═ κ (Mx, Mx'), so the gaussian kernel matrix is a circulant matrix.
(4) Carrying out rapid detection;
f(z)=(Kz)Tα (15)
Kzthe kernel matrix of the training sample and the candidate image area is known according to the cyclic displacement and kernel function theorem:
wherein k isxzIs a nuclear correlation between x and z and is KzThe first row of the matrix. Substituting equation (16) into equation (15) to obtain:
when the value of the coordinate is the maximum value, the corresponding coordinate represents the change of the position of the target area, and the coordinate is added with the coordinate of the target area of the previous frame to obtain the position of the target area of the frame.
(5) Updating the model;
in the target tracking process of the KCF algorithm, a target appearance model needs to be updated in real time, and the updating strategy is as follows:
α=(1-β)αt-1+βαt (20)
x=(1-β)xt-1+βxt (21)
where β is the learning rate of the model, αt-1Representing training samples x from the previous framet-1Training the resulting coefficient vector, αtRepresenting the input sample x from the current new inputtAnd training the obtained coefficient vector, wherein alpha represents the coefficient vector of the next frame, and x represents the training sample of the next frame.
The position of the dial in the video stream is obtained through a KCF tracking algorithm, and the specific positioning method is shown in figure 2.
Firstly, setting the area of the dial in a first frame image, reading the length and the width of the area of the dial and the coordinates of the top left corner of the area, and then calculating the coordinates of the center of the dial; the search area is set to be 2.5 times the area of the dial plate, so that environmental factors and negative samples can be provided for the tracking area. A gaussian regression label y is created. Determining the parameter lambda 10-4δ is 0.5 and β is 0.02. According to the target area (the area of the dial)Domain) xtThe Gaussian kernel autocorrelation is obtained from the formula (4)Is obtained from the formula (14)Let x be xt,
Then inputting a new image sequence; and preliminarily determining the central coordinate of the table in the previous frame of image according to the central coordinate of the table in the previous frame of image. The Gaussian kernel cross-correlation is obtained from the formula (4)Is obtained from the formula (13)Is obtained from the formula (19)The corresponding coordinate when the value is maximum represents the position change of the frame target area relative to the last frame target area. Adding the coordinate of the table center of the previous frame to the coordinate of the table center of the previous frameAnd obtaining the accurate position of the central coordinate of the dial area in the frame image by the corresponding coordinate when the value is maximum. Updating the target area x according to the central coordinate value of the dialtUpdating Gaussian kernel autocorrelationThereby updatingFurther according to the formulas (20) and (21),updating input of next frame imagex. And (5) circulating 1.2 in the way until the image sequence is processed. An example of dial positioning results is shown in fig. 3.
Step 2: correcting dial angle offset;
due to the shaking of the machine, the dial plate in the picture captured by the camera is not horizontally placed. And (3) correcting the angle offset of the dial plate by utilizing Speeded Up Robust Features (SURF for short) in the dial plate area framed in the step (1) to level the dial plate.
A Speeded Up Robust Features (SURF for short) algorithm is proposed by Herbert Bay, firstly, a Hessian matrix is constructed, and all interest points are generated and used for feature extraction; then constructing a scale space; then comparing the pixel points processed by the Hessian matrix with 26 adjacent points in the 3-dimensional space field, if the pixel points are larger or smaller than the other 26 points, reserving the pixel points as primary feature points, filtering out the feature points with weaker energy and wrong positioning, and screening out the final stable feature points; counting harr wavelet characteristics in the neighborhood of the characteristic points and determining the main direction of the characteristic points; generating a feature point descriptor; and finally, determining the matching degree by calculating the Euclidean distance between two feature points, wherein the shorter the Euclidean distance is, the better the two feature points are matched.
And step 3: detecting a pointer;
and (3) preprocessing the image corrected in the step (2), detecting a straight line by using a Hough straight line detection method, and finally selecting the best straight line to represent the pointer.
The idea of detecting straight lines by Hough transformation is as follows: a line is denoted by (r, theta), where r is the distance of the line from the origin and theta is the angle between the perpendicular to the line and the x-axis. A straight line in the x-y coordinate system is represented as a point in the r-theta coordinate system. There are n straight lines through a point in the x-y coordinate system, and then the n straight lines are n points in the r-theta coordinate system. To determine whether m points in the x-y coordinate system are on a straight line, drawing m x n straight lines by a traversal method, correspondingly, m x n (r, theta) coordinates in the r-theta coordinate system, if the theta of the m points is equal to thetaiWhen r is equal to riThen m points are proved to be on a straight line. In the actual line detection case, if more than a certain number of points have the same (r, theta) coordinates, it can be determined that there is a line.
And (3) performing morphological operation and skeletonization on the image processed in the step (2), and then performing Hough transformation to detect a straight line. The lines detected by Hough transformation are multiple, the best line is selected according to two parameters of the length of the line and the distance from the line to the center of the dial, and the line which is long and is close to the center of the dial is the best line to be found. An example of the detection pointer results is shown in fig. 5.
And 4, step 4: reading;
and 4, obtaining a straight line where the pointer is located, then judging which quadrant the pointer is in relative to the center of the dial plate, and finally solving the number indicated by the pointer according to an angle method.
And (4) taking the center of the dial as the origin of coordinates, and judging that the pointer detected in the step (3) is positioned in the quadrant II. The judging method comprises the following steps: selecting a point of the two end points of the pointer farther from the origin (x, y) with origin coordinates of (x)0,y0) If x > x0,y>y0Then the pointer is in the first quadrant, and so on. And then calculating the included angle between the pointer and the horizontal direction. The ratio of the included angle between the pointer and the minimum scale to the total angle is equal to the ratio of the current scale to the measuring range, and therefore the current scale is obtained.
The effect of the invention can be illustrated by the following simulations:
the test data for the first experiment was derived from the factory shot example, as shown in fig. 5. The pointer is fixed at 0.37 position in the shooting picture, and the camera changes the shooting angle, the shooting distance and the shooting light so as to increase the detection difficulty. As can be seen from FIG. 5, when the angle is shifted, the distance is increased, and the light is darkened, the pointer can be accurately detected by the method, the reading error is within 0.0035, and the speed is 0.05s per frame on average.
And the second experiment and the third experiment are dials of pointer movement shot by the mobile phone. The shaking of the machine is simulated, the shooting angle of the mobile phone is changed, and the dial plate in the picture captured by the mobile phone is horizontal or inclined, or clear or fuzzy. The dial plate used in the second experiment is shown in figure 3, the measuring range is 60MPa, the minimum interval is 1MPa, and the error of automatic reading is lower than 0.36 MPa; the dial plate used in the third experiment is shown in the figure 3, the measuring range is 0.4MPa, the minimum interval is 0.01MPa, and the error of automatic reading is lower than 0.003 MPa. The speed averages 0.05s per frame. Table 1 gives part of the experimental data.
The effectiveness and the practicability of the invention are demonstrated by simulation experiments, the time delay of the existing positioning dial plate area algorithm is greatly improved, the bottleneck that the dial plate level in a picture needs to be shot when the reading is carried out is solved, and the practical application requirements are met.
TABLE 1 Dial automatic identification result and observed value comparison (MPa)
It should be understood that parts of the specification not set forth in detail are well within the prior art.
It should be understood that the above description of the preferred embodiments is given for clarity and not for any purpose of limitation, and that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (4)
1. A method for identifying a dial indicator by using computer vision is characterized by comprising the following steps:
step 1: positioning the area of the dial plate;
in a first frame image of a video, framing an area where a dial is located, determining the length and width of the area where the dial is located and coordinates of a top left corner vertex of the area, and solving a central coordinate of the dial; then, tracking the dial by using a KCF algorithm, and determining the area where the dial is located and the center coordinate of the dial in each frame of image;
step 2: correcting dial angle offset;
and step 3: detecting a pointer;
and 4, step 4: and (6) reading.
2. The method for identifying a dial indicator using computer vision according to claim 1, wherein: in step 2, the central coordinates of the area where the dial is located in the image sequence obtained in step 1 are used, and the value with the smaller value of the length and the width of the area where the dial is located is used as the radius of the dial, so that the dial area is determined; firstly, making pixels of a non-dial area in a 1 st frame image be zero, and regarding the image as original to eliminate the interference of the surrounding environment; then, making pixels of a non-dial area in the images from the 2 nd frame to the nth frame be zero, and regarding the image as image; respectively detecting SURF characteristics of original and image and extracting; then carrying out feature matching on the original and the image to find out the position of a feature point in the original image in the image; estimating the offset angles of the two images according to the feature matching result; and finally, compensating the angle deviation to enable the dial to be horizontal.
3. The method for identifying a dial indicator using computer vision according to claim 1, wherein: and 3, performing morphological operation and skeletonization pretreatment on the image corrected in the step 2, detecting straight lines by using a Hough straight line detection method, wherein a plurality of straight lines are detected by Hough transformation, and the optimal straight line is selected according to two parameters of the length of the straight line and the distance from the straight line to the center of the dial plate, so that the straight line which is long and is close to the center of the dial plate is the optimal straight line.
4. The method for identifying a dial indicator using computer vision according to any one of claims 1 to 3, wherein: step 4, regarding the center of the dial as the origin of coordinates, and judging that the pointer detected in the step 3 is positioned in the fourth quadrant; the judging method comprises the following steps: selecting a point of the two end points of the pointer farther from the origin (x,y) The origin coordinate isx 0,y 0) If, ifx>x 0,y>y 0If the pointer is in the first quadrant, and so on; then, an included angle between the pointer and the horizontal direction is calculated, and the ratio of the included angle between the pointer and the minimum scale to the total angle is equal to the ratio of the current scale to the measuring range, so that the current scale is obtained.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710813475.6A CN109359643B (en) | 2017-09-11 | 2017-09-11 | Dial plate pointer identification method using computer vision |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710813475.6A CN109359643B (en) | 2017-09-11 | 2017-09-11 | Dial plate pointer identification method using computer vision |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109359643A CN109359643A (en) | 2019-02-19 |
CN109359643B true CN109359643B (en) | 2022-05-13 |
Family
ID=65349680
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710813475.6A Expired - Fee Related CN109359643B (en) | 2017-09-11 | 2017-09-11 | Dial plate pointer identification method using computer vision |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109359643B (en) |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1122521A1 (en) * | 2000-02-01 | 2001-08-08 | Setrix AG | Method and apparatus for monitoring an analog meter |
CN104899609B (en) * | 2015-06-19 | 2019-03-26 | 四川大学 | A kind of mechanical meter recognition methods based on image registration |
CN105303168A (en) * | 2015-10-14 | 2016-02-03 | 南京第五十五所技术开发有限公司 | Multi-view pointer type instrument identification method and device |
CN105740856B (en) * | 2016-01-28 | 2019-01-18 | 宁波理工监测科技股份有限公司 | A kind of pointer instrument registration read method based on machine vision |
CN106650592B (en) * | 2016-10-05 | 2020-08-28 | 北京深鉴智能科技有限公司 | Target tracking system |
CN106991695A (en) * | 2017-03-27 | 2017-07-28 | 苏州希格玛科技有限公司 | A kind of method for registering images and device |
-
2017
- 2017-09-11 CN CN201710813475.6A patent/CN109359643B/en not_active Expired - Fee Related
Also Published As
Publication number | Publication date |
---|---|
CN109359643A (en) | 2019-02-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107423702B (en) | Video target tracking method based on TLD tracking system | |
CN110580723B (en) | Method for carrying out accurate positioning by utilizing deep learning and computer vision | |
CN108875794B (en) | Image visibility detection method based on transfer learning | |
CN108109162B (en) | Multi-scale target tracking method using self-adaptive feature fusion | |
CN104657711B (en) | A kind of readings of pointer type meters automatic identifying method of robust | |
CN105427298A (en) | Remote sensing image registration method based on anisotropic gradient dimension space | |
CN108121945A (en) | A kind of multi-target detection tracking, electronic equipment and storage medium | |
CN111950396A (en) | Instrument reading neural network identification method | |
CN110956131B (en) | Single-target tracking method, device and system | |
CN110910456B (en) | Three-dimensional camera dynamic calibration method based on Harris angular point mutual information matching | |
CN110659637A (en) | Electric energy meter number and label automatic identification method combining deep neural network and SIFT features | |
CN116229052B (en) | Method for detecting state change of substation equipment based on twin network | |
CN115456956A (en) | Method and device for detecting scratches of liquid crystal display and storage medium | |
CN108596032B (en) | Detection method, device, equipment and medium for fighting behavior in video | |
CN112652020A (en) | Visual SLAM method based on AdaLAM algorithm | |
CN115147418A (en) | Compression training method and device for defect detection model | |
CN106101640A (en) | Adaptive video sensor fusion method and device | |
CN108830828A (en) | A kind of method for detecting change of remote sensing image and device | |
CN114463397A (en) | Multi-modal image registration method based on progressive filtering | |
CN105427276A (en) | Camera detection method based on image local edge characteristics | |
CN109359643B (en) | Dial plate pointer identification method using computer vision | |
Du et al. | Grid-based matching for full-field large-area deformation measurement | |
CN105447524A (en) | Image identification method and device | |
Huang et al. | A checkerboard corner detection method using circular samplers | |
Tong et al. | Surface Defect Detection Method Based on Improved Faster-RCNN |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20220513 |
|
CF01 | Termination of patent right due to non-payment of annual fee |