CN107578433B - Method for identifying temperature of electrode plate of electrolytic cell - Google Patents
Method for identifying temperature of electrode plate of electrolytic cell Download PDFInfo
- Publication number
- CN107578433B CN107578433B CN201710707859.XA CN201710707859A CN107578433B CN 107578433 B CN107578433 B CN 107578433B CN 201710707859 A CN201710707859 A CN 201710707859A CN 107578433 B CN107578433 B CN 107578433B
- Authority
- CN
- China
- Prior art keywords
- image
- edge
- visible light
- electrolytic cell
- temperature
- 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.)
- Active
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P10/00—Technologies related to metal processing
- Y02P10/20—Recycling
Abstract
The invention provides a method for identifying the temperature of an electrolytic cell polar plate, which comprises the following steps: step S1, acquiring a visible light image and a thermal image of the electrolytic cell, and correspondingly registering the visible light image and the thermal image to acquire a spatial position thermal image; step S2, matching the visible light image with standard edge templates of an electrolytic cell and a polar plate respectively to obtain a spatial position image; and step S3, identifying the real-time temperature value of the polar plate based on the spatial position thermal image and the spatial position image. The method for identifying the temperature of the polar plate of the electrolytic cell can effectively improve the accuracy and reliability of the identification of the temperature of the polar plate.
Description
Technical Field
The invention relates to an image processing technology, in particular to a method for identifying the temperature of an electrolytic cell polar plate.
Background
The copper electrolysis process is that the rough copper is made into the latter half in advance and used as an anode, the pure copper is made into a sheet and used as a cathode, and the sheet is inserted into the electrolyte at intervals. After the electricity is supplied, copper is dissolved into copper ions from the anode and moves to the cathode, electrons are obtained after the copper ions reach the cathode, and pure copper is separated out at the cathode.
In copper electrolytic refining, current efficiency and grade rate are important indexes for examining the electrolytic production efficiency and capacity, and the short circuit between electrodes has adverse effects on the two indexes. When the short circuit occurs between the electrodes, on one hand, part of electric energy is consumed due to heating, and the current efficiency naturally also decreases; on the other hand, on the cathode with the inter-electrode short circuit, the inter-electrode short circuit part of the plate surface generally has large-area coarse crystals and grains, so that the high-purity cathode copper cannot be counted by the plate, the grade rate is influenced, and therefore, the timely discovery of the inter-electrode short circuit problem is very important.
At present, the existing detection mode is to use a thermal infrared imager to indirectly judge the short circuit condition by detecting the temperature. The thermal infrared image is a gray image, has no color or shadow, low contrast and fuzzy visual effect; random interference of external environment and imperfection of an infrared imaging system bring various noises to the thermal infrared image, and the noises with complex distribution cause the signal-to-noise ratio of the thermal infrared image to be high, thereby being not beneficial to the processing of target identification in the subsequent link.
The defects of fuzzy edge profile of the electrolytic cell polar plate, poor background contrast ratio and the like generally exist in the thermal infrared image, if the thermal infrared imager is arranged at a longer distance and is influenced by severe conditions of an industrial field, the signal-to-noise ratio and the contrast ratio of the obtained thermal infrared image are lower, the image quality is poor, the positioning error of the electrolytic cell polar plate is large, and the temperature of the polar plate cannot be accurately and reliably identified.
Disclosure of Invention
The invention provides a method for identifying the temperature of an electrolytic cell plate, which overcomes or at least partially solves the problems so as to solve the technical problems of low accuracy and poor contrast of plate temperature identification.
According to one aspect of the invention, there is provided a method for identifying the temperature of an electrolytic cell plate, comprising:
step S1, acquiring a visible light image and an infrared thermal image of the electrolytic cell, and correspondingly registering the visible light image and the infrared thermal image to acquire a spatial position thermal image;
step S2, matching the visible light image with standard edge templates of an electrolytic cell and a polar plate respectively to obtain a spatial position image;
step S3, identifying a real-time temperature of the plate based on the spatial position thermal image and the spatial position image.
Further, still include: and step S4, determining the position of the electrolytic cell corresponding to the visible light image based on the layout conditions of the electrolytic cell and the polar plate.
Further, the correspondingly registering the visible light image and the infrared thermal image in step S1 to acquire a spatial position thermal image specifically includes:
step S11, filtering and denoising the original visible light image and the original infrared thermal image to obtain the visible light image and the infrared thermal image;
and step S12, taking the visible light image as a reference image, and adopting an affine transformation model to realize the registration of the spatial positions of the visible light image and the infrared thermal image.
Further, the registration of the spatial positions of the visible light image and the infrared thermal image in step S12 employs a point-feature-based image matching model, which specifically includes:
step S121, extracting feature points by using a Harris operator, and establishing a feature descriptor for each feature point;
step S122, matching the feature points by taking the absolute value distance between the feature descriptors as similarity measurement;
s123, eliminating error matching points by adopting a random sampling consistency algorithm to obtain a matching point set;
and S124, solving the optimal transformation parameter of the matching point set by adopting a least square method, and realizing the image geometric registration of the visible light image and the infrared thermal image.
Further, the step S2 of matching the visible light pattern with the standard edge templates of the electrolytic cell and the plate specifically includes:
step S21, performing edge detection on the visible light image to obtain a visible light edge binary image;
step S22, constructing an electrolytic bath standard edge template, and matching the electrolytic bath standard edge template with the visible light edge binary image to extract an electrolytic bath edge image in the visible light image;
and S23, constructing a standard plate edge template of the electrolytic cell, and matching the standard plate edge template with the edge map of the electrolytic cell to identify the plate edge map in the electrolytic cell in the visible light image.
Further, the edge detection on the visible light image in step S21 specifically includes:
s211, extracting an edge point set by adopting an edge operator;
and S212, removing the deviation points in the edge point set and carrying out corresponding filling to obtain an optimized edge point set, and connecting the optimized edge point set into a line.
Further, in step S211, the edge operator is a Canny operator, and extracting the edge point set by using the Canny operator specifically includes:
step S2111, carrying out non-maximum suppression processing on the pixel gradient in the visible light image to obtain a non-maximum suppression image;
step S2112, double-threshold processing is carried out on the non-maximum value inhibition image to obtain a low-threshold edge image and a high-threshold edge image so as to extract an edge point set.
Further, the similarity when the visible light image is matched with the standard edge template of the electrolytic cell or the electrode plate in the electrolytic cell is represented by a correlation coefficient R (i, j), and the correlation coefficient R (i, j) is calculated by adopting the following formula:
wherein T (m, n) is the pixel value of the standard edge template of the electrolytic cell or the polar plate, SijAnd (m, n) is the pixel value of the edge image of the electrolytic bath or the edge image of the polar plate.
Further, the step S3 of identifying the real-time temperature of the plate specifically includes: collecting temperature points in the edge range of the polar plate, adopting a Romanov criterion to remove the temperature point with the largest error in the temperature points, obtaining effective temperature points, and determining the temperature of the polar plate according to the average temperature value of the effective temperature points.
Further, in step S4, determining the position of the electrolytic cell corresponding to the visible light image based on the layout conditions of the electrolytic cell and the electrode plate specifically includes:
step S41, establishing a rectangular coordinate system with the width direction of the electrolytic cell array as an x axis and the length direction as a y axis;
step S42, acquiring a fixed x value of the electrolytic cell based on a contact switch arranged on a guide rail of the travelling crane;
s43, determining a y value according to the distance between the travelling crane and the edge of the electrolytic cell array measured by the laser range finder of the travelling crane;
and step S44, determining the position information of the current electrolytic tank based on the x value and the y value, and determining the pixel area corresponding to each polar plate in the electrolytic tank on the infrared thermal image by combining the number of the polar plates in the electrolytic tank.
The beneficial effects of the invention are mainly as follows:
(1) registering the acquired visible light image and the infrared thermal image, so that the temperature information of the electrode plate acquired in the infrared thermal image can be intuitively reflected in the visible light image, and the temperature information of the electrode plate in the electrolytic cell can be more intuitively monitored;
(2) matching the obtained visible light image with the standard edge templates of the corresponding electrolytic cell and the corresponding polar plate, so that the obtained visible light image corresponds to the positions of the electrolytic cell and the polar plate one by one, and the real-time temperature of the corresponding polar plate can be conveniently and accurately identified;
(3) determining the actual position of the electrolytic cell in the electrolytic cell array corresponding to the current visible light image and the infrared thermal image according to the arrangement condition of the electrolytic cell and the polar plate, so as to conveniently and quickly search the corresponding polar plate position;
(4) after the temperature of the polar plate is identified, the temperature point value in the edge range of the polar plate is further processed by adopting the Romanov criterion, so that the influence of measurement errors is reduced, and the temperature of the polar plate is identified more accurately.
Drawings
FIG. 1 is a schematic flow diagram of a method for identifying the temperature of an electrolytic cell plate according to an embodiment of the present invention;
FIG. 2 is a schematic image registration flow chart of a method for identifying the temperature of an electrolytic cell plate according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a standard edge template matching for a method for identifying the temperature of an electrolyzer plate according to an embodiment of the invention;
FIG. 4 is a graph showing the effect of edge detection of a visible light image in a method for identifying the temperature of an electrode plate of an electrolytic cell according to an embodiment of the present invention;
FIG. 5 is a partially enlarged view of an edge detection effect graph of a visible light image of a method for identifying the temperature of an electrode plate of an electrolytic cell according to an embodiment of the invention;
FIG. 6 is a schematic sector view of a method for identifying the temperature of the plates of an electrolyzer according to an embodiment of the invention.
Detailed Description
The following detailed description of embodiments of the present invention is provided in connection with the accompanying drawings and examples. The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention.
Referring to fig. 1, a method for identifying the temperature of an electrolytic cell plate comprises:
s1, acquiring a visible light image and an infrared thermal image of the electrolytic cell, and correspondingly registering the visible light image and the infrared thermal image to acquire an infrared thermal image of a spatial position;
step S2, matching the visible light image with standard edge templates of an electrolytic cell and a polar plate respectively to obtain a spatial position image;
and step S3, identifying the real-time temperature of the polar plate based on the spatial position infrared thermal image and the spatial position image.
The electrolytic cells in the production workshop are arranged in an array form to form an electrolytic cell array. And respectively establishing an electrolytic cell standard edge template and an electrolytic cell inner polar plate standard edge template in the electrolytic cell array according to the actual arrangement position of each electrolytic cell in the electrolytic cell array and the arrangement condition of the electrolytic cell inner polar plate, so as to identify the corresponding positions of the electrolytic cell and the polar plate in the visible light image.
The acquired visible light image and the infrared thermal image are registered, so that the temperature information of the polar plate reflected in the infrared thermal image corresponds to the scene in the visible light image, and the temperature information of the polar plate can be intuitively and clearly reflected in the visible light image.
Meanwhile, after the visible light image of the current electrolytic cell is obtained, the visible light image is sequentially matched with the standard edge templates of the electrolytic cell and the polar plate, so that the actual positions of the electrolytic cell and the polar plate are clearly reflected in the visible light image.
The matched standard edge template and the visible light image as well as the registered visible light image and the infrared thermal image are integrated, so that the temperature information reflected in the infrared thermal image can correspond to the position relationship between the electrolytic cell and the polar plate one by one, and the temperature information of the corresponding polar plate can be identified more accurately and rapidly.
Specifically, a high-resolution camera can be used for obtaining and processing a visible light image, and a thermal infrared imager can be used for obtaining and processing an infrared thermal image. The high-resolution camera and the thermal infrared imager can move in the horizontal direction above the electrolytic cell array so as to scan all electrolytic cells in the whole electrolytic cell array one by one, thereby acquiring visible light images and infrared thermal images of all electrolytic cells and electrode plates in the electrolytic cells. In the invention, the temperature identification method of the polar plate in one electrolytic tank is taken as an example, and the temperature identification methods of the polar plates of other electrolytic tanks are the same.
In a specific embodiment, the method further comprises step S4 of determining the position of the electrolytic cell corresponding to the visible light image based on the layout conditions of the electrolytic cell and the electrode plate. After the temperature information in the infrared thermal image is associated with the positions of the electrolytic cells and the plates in the visible light image, the correspondence between the current visible light image and the actual positions of the electrolytic cells in the electrolytic cell array is further determined.
In another specific embodiment, the correspondingly registering the visible light image and the infrared thermal image in step S1 to obtain the spatial position infrared thermal image specifically includes:
step S11, filtering and denoising the original visible light image and the original infrared thermal image to obtain the visible light image and the infrared thermal image;
and step S12, taking the visible light image as a reference image, and adopting an affine transformation model to realize the registration of the spatial positions of the visible light image and the infrared thermal image.
After a visible light image and an infrared thermal image of the electrolytic cell are obtained, the visible light image is taken as a reference image, the infrared thermal image is taken as an image to be registered, a corresponding common object in the image is utilized, a relative position mapping relation between the visible light image and the infrared thermal image is found out through comparison and matching, the spatial position and the image gray level are aligned, and finally a corresponding relation between the visible light image and the infrared thermal image is established, so that the temperature information embodied in the infrared thermal image is correspondingly associated with the position information of the electrolytic cell and/or the polar plate in the visible light image, the temperature of the polar plate is identified and extracted, and the real-time and accurate monitoring of the temperature of the polar plate is realized.
The original visible light image is directly shot by a camera, and the original infrared thermal image is directly obtained by a thermal infrared imager. After the original visible light image and the original infrared thermal image are obtained, filtering and denoising processing is firstly carried out on the original visible light image and the original infrared thermal image so as to obtain the visible light image and the infrared thermal image. The filtering noise reduction process may employ a gaussian low pass filter.
The gaussian low-pass filter is a linear smoothing filter with a transfer function of a gaussian function, and since the gaussian function is a density function of a normal distribution, the gaussian low-pass filter is very effective in removing noise that follows a normal distribution (Normaldistribution). Since the acquired original visible light image and original infrared thermal image are two-dimensional signals, the two-dimensional gaussian function is used as a transfer function for the image filtering and noise reduction processing, and is expressed as follows:
in the formula, δ is a standard deviation.
The gaussian function has separable characteristics, so that the gaussian filter is performed on the columns before the gaussian filter is performed on the columns to simplify the processing. After the above processing, the two-dimensional gaussian function can be reduced to a one-dimensional gaussian function such as:
in the formula, δ is a standard deviation.
In the actual registration process, the visible light image and the infrared thermal image can be represented as two-dimensional matrices, respectively: visible light image I1Infrared thermal image I2。I1(x,y)、I2(x, y) respectively represent the gray values of the two images at the point (x, y), and then the visible light image I1And infrared thermal image I2The registration relationship of (a) can be expressed by the following transformation formula:
I1(x,y)＝g(I2(f(x,y))) (3)
wherein, I1(x,y)、I2(x, y) respectively represent the gray values of the two images at the point (x, y), f represents a two-dimensional space geometric transformation function, and g represents a one-dimensional gray transformation function.
During the registration process of the visible light image and the infrared thermal image, the alignment between the spatial position and the gray scale of the corresponding pixel of the image can be embodied, and the alignment of the spatial position can be embodied by the solution of the transformation formula (3) to the function f; the alignment in the image gray scale can be reflected in the solution of the function g by the transformation equation (3).
In the process of identifying the temperature of the electrolytic cell polar plate, the geometric position of the infrared thermal image corresponding to the visible light image is changed, and the temperature information of the electrolytic cell polar plate can be accurately and visually acquired. Therefore, to simplify the operation process and reduce the operation amount, we only find the space geometric transformation relation f (x, y), and the above expression can be rewritten into a more general expression:
I1(x，y)＝I2(f(x，y)) (4)
an affine transformation is used for the spatial geometric transformation model during the registration of the visible light image and the infrared thermal image in the spatial position. The affine transformation is composed of Cartesian transformation of scale, translation and rotation, and can effectively match two images of the same scene with the same visual angle and the changed image acquisition positions. Because the spatial geometric relationship between the point sets is not changed by the transformation process when the affine transformation is adopted, the affine transformation belongs to the global rigid body transformation.
Affine transformations typically have six parameters: a is11、a12、a21、a22、tx、ty. Under the affine transformation model, let (x)1,y1)、(x2,y2) Two corresponding points in the reference image and the image to be registered are respectively, and the corresponding relationship between the two points can be expressed as follows:
the straight line on the image to be registered after affine transformation is mapped into the reference image and still is a straight line, the balance relation is kept, the method is more universal than rigid body transformation, besides translation and rotation, inclination and aspect ratio change in distortion are considered, and the registration accuracy of the visible light image and the infrared thermal image on the spatial position can be improved.
Referring to fig. 2, in another specific embodiment, since the difference between the gray scale information of the visible light image and the gray scale information of the infrared thermal image is relatively large, in the registration process, the visible light image and the infrared thermal image adopt image matching based on point features to improve the registration accuracy. In step S12, the registration of the spatial positions of the visible light image and the infrared thermal image uses an image matching model based on point features, and the image matching model based on point features specifically includes:
step S121, extracting feature points by using a Harris operator, and establishing a feature descriptor for each feature point;
step S122, matching the feature points by taking the absolute value distance between the feature descriptors as similarity measurement;
s123, eliminating error matching points by adopting a random sampling consistency algorithm to obtain a matching point set;
and S124, solving the optimal transformation parameter of the matching point set by adopting a least square method, and realizing the image geometric registration of the visible light image and the infrared thermal image.
By adopting the image matching based on the point characteristics, the calculated amount in the registration process can be reduced to a great extent, and the registration accuracy is improved.
In the registration process, corresponding Gaussian pyramid models are respectively established for the visible light image and the infrared thermal image, the Harris operator is adopted to respectively extract the corner points of each layer of image in the visible light image and the infrared thermal image as feature points, the gradient direction histogram in the smaller neighborhood of each feature point is counted, the direction corresponding to the main peak value of the histogram is taken as the main direction of the feature point, and the Harris autocorrelation matrix trace is used for establishing a feature descriptor for each feature point.
Specifically, the characteristic descriptors of the point A and the point B are PAAnd PBThen the absolute distance between their feature descriptors
Wherein, PA＝[a1a2…an]，PB＝[b1b2…bn]。
The specific matching method comprises the following steps: from visible light image I1Taking feature point A, and obtaining infrared thermal image I2Finding out the point B with the nearest characteristic descriptor of the point A and the next nearest point C, if the point A isAnd the characteristic descriptors of the point B and the point C satisfy the following distance relationship, and the point A and the point B are considered as initial matching point pairs. The distance relationship can be expressed as follows:
dist(PA,PB)/dist(PA,PC)＜t (6)
wherein, PA、PB、PCRespectively, the feature descriptors corresponding to the point A, the point B and the point C, and t is a proportional threshold value of the ratio of absolute value distances between the feature descriptors of the point B and the point C.
Traverse image I1A set of initial matching point sets corresponding to the visible light image and the infrared thermal image can be obtained. The affine transformation is used to achieve registration between the light image and the infrared thermal image only 3 pairs of points are needed to calculate the 6 parameters required for the transformation equation. In the actual calculation process, a plurality of points can be selected, the optimal transformation parameters are obtained by using a least square method, and then the infrared thermal image is subjected to registration transformation to the visible light image serving as a reference image by adopting an affine transformation model.
In another specific embodiment, the step of matching the acquired visible light image with the standard edge templates of the electrolytic cell and the plate in step S2 specifically comprises:
step S21, performing edge detection on the visible light image to obtain a visible light edge binary image;
step S22, constructing an electrolytic bath standard edge template, and matching the electrolytic bath standard edge template with the visible light edge binary image to extract an electrolytic bath edge image in the visible light image;
and S23, constructing a standard plate edge template of the electrolytic cell, and matching the standard plate edge template with the edge map of the electrolytic cell to identify the plate edge map in the electrolytic cell in the visible light image.
And carrying out edge detection on the visible light image to highlight the local edge of the visible light image and extracting an edge point set of the visible light image so as to obtain a visible light edge binary image expressing the edge profile of the electrolytic cell.
And matching the visible light edge binary image with the constructed standard edge template of the electrolytic cell to extract the edge of the electrolytic cell.
Referring to FIG. 3, for example, a visible edge binary image S obtained by the edge detection process1(W, H), manually constructing standard edge templates T for the electrolytic cell1(m, n). Will S1(W, H) and T1(m, n) matching, identifying the electrolytic cell portion in the visible edge binary map
Judging which edge of the electrolytic cell the center of the visible light image falls in, and extracting the edge of the electrolytic cell as an electrolytic cell edge image S2(W,H)。
The extracted edge graph S of the electrolytic cell2(W, H) and the standard edge template T of the constructed polar plate2(m, n) are matched to identify the plate edge map of each plate in the cell
After identifying the plate edge map, the identified plate edge map is retained in the visible light image to display the position of the plate in the visible light image.
In another specific embodiment, the edge detection on the visible light image in step S21 specifically includes:
s211, extracting an edge point set by adopting an edge operator;
and S212, removing the deviation points in the edge point set and carrying out corresponding filling to obtain an optimized edge point set, and connecting the optimized edge point set into a line.
When the edge detection is carried out on the visible light image, the edge enhancement operator is firstly utilized to highlight the local edge in the visible light image, and then the edge point set is extracted by defining the edge intensity of the pixel in the visible light image and setting the threshold value. However, there may be a case where the extracted edge point set is discontinuous due to the influence of noise and image blurring factors. Therefore, the extracted edge point set needs to be further processed, some deviation points are removed and filled correspondingly, so as to obtain the required visible light edge binary image. The effect of edge detection on visible light images is illustrated in fig. 4-5.
In another specific embodiment, the edge operator in step S211 is a Canny operator, and the extracting the edge point set using the Canny operator specifically includes:
step S2111, carrying out non-maximum suppression processing on the pixel gradient in the visible light image to obtain a non-maximum suppression image;
step S2112, double-threshold processing is carried out on the non-maximum value inhibition image to obtain a low-threshold edge image and a high-threshold edge image so as to extract an edge point set.
Specifically, when the edge enhancement operator is adopted to highlight the local edge in the visible light image, the Canny operator is adopted to carry out edge detection on the visible light image so as to extract the edge point set. And filtering the original visible light image by adopting a Gaussian filter to remove noise in the visible light image to obtain the visible light image. For each pixel in the visible light image, the magnitude M (i, j) and direction θ (i, j) of its gradient are calculated. The first order approximation P (i, j) of partial differential of each pixel gradient in the x-direction and the first order approximation Q (i, j) of partial differential in the y-direction are:
P(i,j)＝(S(i,j+1)-S(i,j)+S(i+1,j+1)-S(i+1,j))/2 (7)
Q(i,j)＝(S(i,j)-S(i+1,j)+S(i,j+1)-S(i+1,j+1))/2 (8)
in the formula, S (i, j), S (i, j +1), S (i +1, j), and S (i +1, j +1) are visible light image pixel values.
From this, the magnitude M (i, j) and direction θ (i, j) of the pixel gradient can be calculated.
Where P (i, j) is a first order approximation of the partial differential of the pixel gradient in the x-direction and Q (i, j) is a first order approximation of the partial differential of the pixel gradient in the y-direction.
θ(i,j)＝arctan[P(i,j)/Q(i,j)](10)
Where P (i, j) is a first order approximation of the partial differential of the pixel gradient in the x-direction and Q (i, j) is a first order approximation of the partial differential of the pixel gradient in the y-direction.
Referring to fig. 6, the range of variation of the gradient angle θ (i, j) is narrowed to one of four sectors numbered 0 to 3 corresponding to four possible combinations of elements in a 3 × 3 neighborhood, a 3 × 3 template is used to act on all points in the array of magnitudes of the pixel gradient, and the Sector value ξ (i, j) at the center in the neighborhood is set to Sector (θ (i, j)).
At each point, the center pixel of the neighborhood is compared to two elements along a gradient line given by the sector value ξ (i, j) at the center of the neighborhood, M (i, j) is assigned zero if the magnitude M (i, j) at the center point of the neighborhood is not greater than the magnitude of two adjacent points along the gradient line.
After the pixel gradient is subjected to non-maximum suppression processing, the image N (i, j) obtained by the non-maximum suppression processing is processed in a dual-threshold mode, namely two thresholds tau are used1And τ2The images N (i, j) are processed separately. Of these, two thresholds τ are preferred1And τ2Is τ2＝2·τ1。
Setting the gradient value less than tau1The gray scale of the pixel (2) is set to 0, and a low threshold edge image T is obtained1In the same way, the high-threshold edge image T can be obtained2. Edge image T due to high threshold2Is obtained using a high threshold, and therefore, it contains fewer false edges, while at the same time some useful edge information is lost; low threshold edge image T1Is derived from the low threshold, it is able to retain more edge information, but at the same time, there are more false edges.
After extracting the edge point set of the visible light image by using Canny operator, using the acquired low threshold edge image T1And high threshold edge image T2And on the basis, removing the deviation points and filling correspondingly, and then integrally connecting the processed edge points to form a line to obtain the visible light edge binary image.
Specifically, the high threshold edge image T is scanned2When a non-zero gray pixel P is encountered, the contour line starting at P is traced until the end point Q of the line. Then at the low threshold edge image T1Middle comparison and high threshold edge image T2Eight adjacent areas of the Q' point corresponding to the middle Q point position. If a pixel R 'with non-zero gray is present in the eight neighbourhood of the point Q', it is included in the image T2And (4) as point R. Similarly, repeat at high threshold edge image T2To continue to find a contour line whose tracking starts at R, and so on until the image T is at the low threshold edge1And high threshold edge image T2Can not be continued until. The connection of the contour containing P has been completed and marked as visited. Then, each contour line in the image can be repeatedly searched in turn until the image T is at the high threshold edge2Until no new contour can be found.
In another specific embodiment, when the visible light image is matched with the standard edge templates of the electrolytic cell and the polar plate respectively, the similarity of the matching degree can be represented by a correlation coefficient R (i, j), and the correlation coefficient can be calculated by the following formula:
wherein T (m, n) is the pixel value of the standard edge template of the electrolytic cell or the polar plate, SijAnd (m, n) is the pixel value of the edge image of the electrolytic bath or the edge image of the polar plate.
It can be understood that, because the standard edge template of the electrolytic bath or the electrode plate and the electrolytic bath/electrode plate edge map are binary maps or images after binarization processing, the pixel value has only 0 or 1 change.
The correlation coefficient is obtained by normalization processing of a relational expression for measuring the similarity of the correlation coefficient, and the similarity relational expression is expressed as follows:
wherein T (m, n) is the mark of the electrolytic cell or the polar platePixel value of quasi-edge template, SijAnd (m, n) is the pixel value of the edge image of the electrolytic bath or the edge image of the polar plate.
When the standard edge template of the electrolytic cell or the polar plate is identical with the matched edge map, the correlation coefficient R (i, j) is 1. After all searches are completed from left to right and from top to bottom in the edge map S of the searched electrolytic cell or plate, the maximum value R of R (i, j) is foundmax(im,jm) Its corresponding edge map S (i)m,jm) The target is obtained. When a plurality of targets are identified simultaneously, a threshold value of R needs to be set, and all targets larger than the threshold value are stored.
Specifically, after the visible light image and the infrared thermal image are registered and the optical image is matched with the standard edge templates of the electrolytic cell and the polar plate, all the electrolytic cells and the corresponding polar plates are correspondingly numbered uniformly according to the actual layout condition of the electrolytic cell array and the polar plates in the electrolytic cell, so that the identified polar plate temperature and the corresponding positions can be conveniently and quickly corresponding.
Specifically, the data is subjected to error analysis through temperature points inside the edge of each polar plate so as to accurately determine the actual temperature value of the polar plate. In another specific embodiment, the step of identifying the real-time temperature of the plate in S3 specifically includes: collecting temperature points in the edge range of the polar plate, adopting a Romanov criterion to remove the temperature point with the largest error in the temperature points, obtaining effective temperature points, and determining the temperature of the polar plate according to the average temperature value of the effective temperature points.
Checking a gross error by adopting a Romanofsky criterion, calculating the error between the average value of the collected plate temperature point values and each temperature point value, removing the temperature point value with the largest error, and obtaining an optimized temperature point value; the t-profile is then used to check whether the optimized temperature point values still contain gross errors. If the gross error still exists, the gross error check is repeated until all temperature point values do not contain the gross error, and the effective temperature point is obtained.
And calculating the average value of all the effective temperature point values to obtain an average temperature value, wherein the average temperature value is the identification value of the plate temperature. Furthermore, the average temperature value obtained by calculation can be used for more accurately reflecting the actual temperature of the electrode plate in a manner of setting a temperature compensation value.
In another specific embodiment, a rectangular coordinate system is established on the arrangement plane of the electrolytic cell array according to the actual arrangement condition of the electrolytic cells in the electrolytic cell array, so as to quickly reflect the actual positions of the current electrolytic cells.
Specifically, the method for acquiring the real-time position of the current electrolytic cell corresponding to the acquired visible light image and the acquired infrared thermal image specifically comprises the following steps:
step S41, establishing a rectangular coordinate system with the width direction of the electrolytic cell array as an x axis and the length direction as a y axis;
step S42, acquiring the x value of the electrolytic cell based on a contact switch arranged on a guide rail of the travelling crane;
step S43, determining a y value based on the distance between the travelling crane and the edge of the electrolytic cell array measured by the laser range finder of the travelling crane;
and step S44, determining the position information of the current electrolytic tank based on the x value and the y value, and determining the pixel area corresponding to each polar plate in the electrolytic tank on the infrared thermal image by combining the number of the polar plates in the electrolytic tank.
A travelling crane is arranged above the electrolytic cell array, the travelling crane is arranged along the width direction of the electrolytic cell array, and the camera and the thermal infrared imager are arranged on an auxiliary crane of the travelling crane. The auxiliary crane moves along the guide rail of the travelling crane to drive the camera and the thermal infrared imager to move along the width direction of the electrolytic cell array.
Specifically, a rectangular coordinate system is established on the plane of the electrolytic cell array so as to determine the current positions of the camera and the thermal infrared imager. The width or length direction of the electrolytic cell array can be arbitrarily taken as an x-axis, and the other direction is taken as a y-axis correspondingly. In a specific embodiment, the width direction of the electrolytic cell array is taken as an x-axis, and the length direction is taken as a y-axis.
The guide rail is provided with a plurality of contact switches, the number of the contact switches is determined according to the number of the electrolytic cells in the row, so that the number of the electrolytic cells arranged in the width direction of the electrolytic cell array corresponds to the number of the contact switches one by one. Preferably, the contact switch is arranged at a position corresponding to the central position of the electrolytic bath, so that the camera and the thermal infrared imager can conveniently acquire complete visible light images and infrared thermal images. That is, the position of the contact switch is used to determine the value of x.
When the auxiliary crane triggers the contact switch and stops moving, the camera and the thermal infrared imager are fixed at one of the contact switches, and the x value is not changed temporarily. And then the travelling crane drives the camera and the thermal infrared imager to move along the y-axis direction together so as to successively acquire visible light images and infrared thermal images of all the electrolytic cells in the row.
In order to conveniently determine the coordinates of the current positions of the camera and the thermal infrared imager in the rectangular coordinate system, the y values of the camera and the thermal infrared imager in the rectangular coordinate system are determined by a laser range finder of the traveling crane. And obtaining the information of the current electrolytic cell opposite to the camera and the thermal infrared imager according to the coordinate x value and the coordinate y value of the camera and the thermal infrared imager.
Preferably, the electrolytic cells in the electrolytic cell array are numbered sequentially, and the length and width data of the electrolytic cells are combined with the coordinates in the rectangular coordinate system, so that the number of the current electrolytic cell can be obtained, the position of the electrolytic cell with abnormal conditions can be conveniently and rapidly positioned, and the electrolytic cell can be conveniently and timely processed. It will be appreciated that the acquisition of the position of the electrolyzer may be carried out simultaneously with the acquisition of the visible light images or infrared thermal images, so as to facilitate a rapid processing of the data.
The invention relates to a method for identifying the temperature of an electrolytic cell polar plate, which is characterized in that the acquired visible light image and the infrared thermal image are registered, the visible light image is matched with the standard edge template of the electrolytic cell and the polar plate, and based on the registered and matched images, the temperature information reflected in the infrared thermal image can be in one-to-one correspondence with the position relation of the electrolytic cell and the polar plate, so that the temperature information of the corresponding polar plate can be identified more accurately and quickly.
Finally, the method of the present invention is only a preferred embodiment and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (9)
1. A method for identifying the temperature of an electrolytic cell plate, comprising:
step S1, acquiring a visible light image and an infrared thermal image of the electrolytic cell, and correspondingly registering the visible light image and the infrared thermal image to acquire a spatial position thermal image;
step S2, matching the visible light image with standard edge templates of an electrolytic cell and a polar plate respectively to obtain a spatial position image;
step S3, identifying the real-time temperature of the pole plate based on the spatial position thermal image and the spatial position image;
the step S3 of recognizing the real-time temperature of the plate specifically includes: collecting temperature points in the edge range of the polar plate, adopting a Romanov criterion to remove the temperature point with the largest error in the temperature points, obtaining effective temperature points, and determining the temperature of the polar plate according to the average temperature value of the effective temperature points.
2. A method for identifying the temperature of an electrolytic cell plate as claimed in claim 1, further comprising: and step S4, determining the position of the electrolytic cell corresponding to the visible light image based on the layout conditions of the electrolytic cell and the polar plate.
3. The method according to claim 1, wherein the step S1 of correspondingly registering the visible light image and the infrared thermal image to obtain the spatial position thermal image specifically comprises:
step S11, filtering and denoising the original visible light image and the original infrared thermal image to obtain the visible light image and the infrared thermal image;
and step S12, taking the visible light image as a reference image, and adopting an affine transformation model to realize the registration of the spatial positions of the visible light image and the infrared thermal image.
4. The method for identifying the temperature of an electrolytic cell plate according to claim 3, wherein the registration of the spatial positions of the visible light image and the infrared thermal image in step S12 employs an image matching model based on point features, which specifically comprises:
step S121, extracting feature points by using a Harris operator, and establishing a feature descriptor for each feature point;
step S122, matching the feature points by taking the absolute value distance between the feature descriptors as similarity measurement;
s123, eliminating error matching points by adopting a random sampling consistency algorithm to obtain a matching point set;
and S124, solving the optimal transformation parameter of the matching point set by adopting a least square method, and realizing the image geometric registration of the visible light image and the infrared thermal image.
5. A method for identifying the temperature of an electrolytic cell plate as claimed in claim 3, wherein matching the visible light pattern with a standard edge template of the electrolytic cell and plate in step S2 includes:
step S21, performing edge detection on the visible light image to obtain a visible light edge binary image;
step S22, constructing an electrolytic bath standard edge template, and matching the electrolytic bath standard edge template with the visible light edge binary image to extract an electrolytic bath edge image in the visible light image;
and S23, constructing a standard plate edge template of the electrolytic cell, and matching the standard plate edge template with the edge map of the electrolytic cell to identify the plate edge map in the electrolytic cell in the visible light image.
6. The method for identifying the temperature of an electrolytic cell plate as claimed in claim 5, wherein the edge detection of the visible light image in step S21 specifically comprises:
s211, extracting an edge point set by adopting an edge operator;
and S212, removing the deviation points in the edge point set and carrying out corresponding filling to obtain an optimized edge point set, and connecting the optimized edge point set into a line.
7. The method according to claim 6, wherein the edge operator in step S211 is a Canny operator, and the extracting the edge point set using the Canny operator specifically comprises:
step S2111, carrying out non-maximum suppression processing on the pixel gradient in the visible light image to obtain a non-maximum suppression image;
step S2112, double-threshold processing is carried out on the non-maximum value inhibition image to obtain a low-threshold edge image and a high-threshold edge image so as to extract an edge point set.
8. The method for identifying the temperature of an electrolytic cell plate as claimed in claim 5, wherein the similarity when the visible light image is matched with a standard edge template of the electrolytic cell or an electrode plate in the electrolytic cell is represented by a correlation coefficient R (i, j), and the correlation coefficient R (i, j) is calculated by the following formula:
wherein T (m, n) is the pixel value of the standard edge template of the electrolytic cell or the polar plate, SijAnd (m, n) is the pixel value of the edge image of the electrolytic bath or the edge image of the polar plate.
9. The method for identifying the temperature of the electrode plate of the electrolytic cell as claimed in claim 2, wherein the step S4 of determining the position of the electrolytic cell corresponding to the visible light image based on the arrangement conditions of the electrolytic cell and the electrode plate specifically comprises:
step S41, establishing a rectangular coordinate system with the width direction of the electrolytic cell array as an x axis and the length direction as a y axis;
step S42, acquiring a fixed x value of the electrolytic cell based on a contact switch arranged on a guide rail of the travelling crane;
step S43, determining a y value based on the distance between the travelling crane and the edge of the electrolytic cell array measured by the laser range finder of the travelling crane;
and step S44, determining the position information of the current electrolytic tank based on the x value and the y value, and determining the pixel area corresponding to each polar plate in the electrolytic tank on the infrared thermal image by combining the number of the polar plates in the electrolytic tank.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710707859.XA CN107578433B (en) | 2017-08-17 | 2017-08-17 | Method for identifying temperature of electrode plate of electrolytic cell |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710707859.XA CN107578433B (en) | 2017-08-17 | 2017-08-17 | Method for identifying temperature of electrode plate of electrolytic cell |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107578433A CN107578433A (en) | 2018-01-12 |
| CN107578433B true CN107578433B (en) | 2020-04-21 |
Family
ID=61035249
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710707859.XA Active CN107578433B (en) | 2017-08-17 | 2017-08-17 | Method for identifying temperature of electrode plate of electrolytic cell |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107578433B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108492335B (en) * | 2018-03-27 | 2020-10-20 | 长春理工大学 | Method and system for correcting perspective distortion of double cameras |
| CN109697433A (en) * | 2019-01-02 | 2019-04-30 | 中南大学 | A kind of aluminium cell degree of superheat state identification method and system based on 3D convolutional neural networks |
| CN111321429B (en) * | 2019-05-13 | 2021-09-21 | 杭州海康威视系统技术有限公司 | Method and device for detecting temperature of electrode plate of electrolytic cell |
| CN113554601A (en) * | 2021-06-30 | 2021-10-26 | 合肥金星机电科技发展有限公司 | Positioning method of electrolytic cell polar plate |
| CN113344908A (en) * | 2021-06-30 | 2021-09-03 | 合肥金星机电科技发展有限公司 | Infrared image validity judgment method for electrolytic cell polar plate detection |
| CN115115630B (en) * | 2022-08-29 | 2022-12-09 | 合肥金星智控科技股份有限公司 | Detection method, detection device, electronic device, and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102567979A (en) * | 2012-01-20 | 2012-07-11 | 南京航空航天大学 | Vehicle-mounted infrared night vision system and multi-source images fusing method thereof |
| CN103712560A (en) * | 2013-12-27 | 2014-04-09 | 镇江冈山电子有限公司 | Part detection method, system and device based on information fusion of multiple sensors |
| CN104463080A (en) * | 2013-09-16 | 2015-03-25 | 展讯通信(天津)有限公司 | Detection method of human eye state |
| CN205643194U (en) * | 2016-05-27 | 2016-10-12 | 南京点触智能科技有限公司 | Touch -control ITO film conducting layer defect detection system based on machine vision |
| CN107025648A (en) * | 2017-03-20 | 2017-08-08 | 中国人民解放军空军工程大学 | A kind of board failure infrared image automatic testing method |
-
2017
- 2017-08-17 CN CN201710707859.XA patent/CN107578433B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102567979A (en) * | 2012-01-20 | 2012-07-11 | 南京航空航天大学 | Vehicle-mounted infrared night vision system and multi-source images fusing method thereof |
| CN104463080A (en) * | 2013-09-16 | 2015-03-25 | 展讯通信(天津)有限公司 | Detection method of human eye state |
| CN103712560A (en) * | 2013-12-27 | 2014-04-09 | 镇江冈山电子有限公司 | Part detection method, system and device based on information fusion of multiple sensors |
| CN205643194U (en) * | 2016-05-27 | 2016-10-12 | 南京点触智能科技有限公司 | Touch -control ITO film conducting layer defect detection system based on machine vision |
| CN107025648A (en) * | 2017-03-20 | 2017-08-08 | 中国人民解放军空军工程大学 | A kind of board failure infrared image automatic testing method |
Non-Patent Citations (3)
| Title |
|---|
| 一种基于Harris算子的快速图像匹配算法;唐永鹤 等;《武汉大学学报 信息科学版》;20120405;第37卷(第4期);第2节 * |
| 基于图像处理技术的电解槽故障诊断方法研究;郭彩乔;《中国优秀硕士学位论文全文数据库 信息科技辑》;20150815;正文第4.1节、第4.3.2节 * |
| 基于红外热像仪的电解槽极板故障在线监测系统;徐勇 等;《安徽科技》;20150720;43-45 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107578433A (en) | 2018-01-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107578433B (en) | Method for identifying temperature of electrode plate of electrolytic cell | |
| CN108107444B (en) | Transformer substation foreign matter identification method based on laser data | |
| CN104574393B (en) | A kind of three-dimensional pavement crack pattern picture generates system and method | |
| CN109658398B (en) | Part surface defect identification and evaluation method based on three-dimensional measurement point cloud | |
| CN108230237B (en) | Multispectral image reconstruction method for electrical equipment online detection | |
| CN110443836A (en) | A kind of point cloud data autoegistration method and device based on plane characteristic | |
| CN105335973B (en) | Apply to the visual processing method of strip machining production line | |
| CN108520514B (en) | Consistency detection method for electronic elements of printed circuit board based on computer vision | |
| CN111640158A (en) | End-to-end camera based on corresponding mask and laser radar external reference calibration method | |
| CN111723821A (en) | Detection and identification method and device for power plant instrument image | |
| CN112241964B (en) | Light strip center extraction method for line structured light non-contact measurement | |
| CN108320310B (en) | Image sequence-based space target three-dimensional attitude estimation method | |
| CN112288682A (en) | Electric power equipment defect positioning method based on image registration | |
| CN113705564B (en) | Pointer type instrument identification reading method | |
| CN113379824B (en) | Quasi-circular fruit longitudinal and transverse diameter measuring method based on double-view-point cloud registration | |
| CN113781413A (en) | Electrolytic capacitor positioning method based on Hough gradient method | |
| CN110334727B (en) | Intelligent matching detection method for tunnel cracks | |
| CN117173590A (en) | Water body abnormality monitoring method based on multisource time sequence remote sensing image | |
| CN109886988B (en) | Method, system, device and medium for measuring positioning error of microwave imager | |
| WO2023134251A1 (en) | Clustering-based light bar extraction method and apparatus | |
| CN109815966A (en) | A kind of mobile robot visual odometer implementation method based on improvement SIFT algorithm | |
| CN112330667B (en) | Morphology-based laser stripe center line extraction method | |
| CN115641326A (en) | Sub-pixel size detection method and system for ceramic antenna PIN needle image | |
| Ziqiang et al. | Research of the algorithm calculating the length of bridge crack based on stereo vision | |
| CN111473944B (en) | PIV data correction method and device for observing complex wall surface in flow field |
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 |