CN115471503A - Equipment abnormity detection method for numerical control ingot splitting machine - Google Patents
Equipment abnormity detection method for numerical control ingot splitting machine Download PDFInfo
- Publication number
- CN115471503A CN115471503A CN202211365049.8A CN202211365049A CN115471503A CN 115471503 A CN115471503 A CN 115471503A CN 202211365049 A CN202211365049 A CN 202211365049A CN 115471503 A CN115471503 A CN 115471503A
- Authority
- CN
- China
- Prior art keywords
- pixel point
- window
- probability
- gray
- acquiring
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 24
- 238000001914 filtration Methods 0.000 claims abstract description 58
- 238000000034 method Methods 0.000 claims abstract description 22
- 230000002159 abnormal effect Effects 0.000 claims abstract description 12
- 239000013598 vector Substances 0.000 claims description 14
- 230000005856 abnormality Effects 0.000 claims description 9
- 230000000694 effects Effects 0.000 abstract description 8
- 238000004364 calculation method Methods 0.000 description 6
- 150000001875 compounds Chemical class 0.000 description 6
- 238000005520 cutting process Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000007405 data analysis Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000017525 heat dissipation Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- DUGOZIWVEXMGBE-UHFFFAOYSA-N Methylphenidate Chemical compound C=1C=CC=CC=1C(C(=O)OC)C1CCCCN1 DUGOZIWVEXMGBE-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000001931 thermography Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10048—Infrared image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- Image Processing (AREA)
- Photometry And Measurement Of Optical Pulse Characteristics (AREA)
Abstract
The invention discloses a method for detecting equipment abnormity of a numerical control ingot splitting machine, belonging to the technical field of data processing; the method comprises the following steps: acquiring a gray scale image of an infrared thermal image of the numerical control ingot slicing machine; acquiring the comprehensive probability of each pixel point being a noise pixel point; acquiring the probability that each pixel point is an edge pixel point; obtaining the size of a filtering window of each pixel point; carrying out mean value filtering on each pixel point according to the size of a filtering window of each pixel point to obtain a denoised gray scale image; and judging whether the equipment of the numerical control ingot splitting machine is abnormal or not according to the average value of the gray scale of the denoised gray scale image. According to the method, the denoising effect of the image is improved by improving the mean filtering algorithm, so that the denoising effect is improved while the image edge is protected; therefore, the accuracy of the equipment abnormity detection of the numerical control ingot splitting machine can be improved.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a method for detecting equipment abnormity of a numerical control ingot splitting machine.
Background
The numerical control ingot slicing machine drives the machine tool to move through a digital program to process a workpiece, long-time operation is needed, and whether the operation of the equipment is normal or not and whether the state is stable or not need to be monitored in real time in order to ensure the safe operation of the equipment. When the traditional method is used for equipment abnormity detection, a large amount of labor cost and financial cost are consumed, on the basis, an equipment abnormity detection method based on intelligent image processing is provided, the running state of equipment is monitored in real time by using an infrared thermal imaging technology, and because the heat production and heat dissipation processes of a machine body are usually kept in dynamic balance when the machine body is in normal running operation, whether the inside of the running equipment is abnormal or not can be detected by analyzing the temperature change of the running equipment.
In order to achieve the purpose, a person skilled in the art judges whether the equipment is abnormal or not through the infrared thermal image in the detection process through equipment abnormality detection of intelligent image processing, but the infrared thermal image is low in contrast and greatly influenced by noise, so that the acquired low-quality infrared thermal image is directly analyzed, and the reliability of an equipment abnormality detection result cannot be guaranteed. The infrared thermal image needs to be denoised to improve the quality of the infrared thermal image. The traditional image denoising methods have respective inherent defects, such as mean value filtering, the size of a filtering window is fixed, the size is selected by experience, image details cannot be well protected when the size of the window is large, and the noise removal effect is poor when the size of the window is small. When the average value is processed, the weights of all the pixel points in the window are the same, so that the influence of the noise pixel points on the average value result is large, and the noise smoothing effect is poor.
Disclosure of Invention
In order to solve the defects in the prior art, the invention provides an equipment anomaly detection method for a numerical control ingot slicing machine, which improves the denoising effect of an image by improving a mean value filtering algorithm, realizes the self-adaptive selection of a window according to the probability that each pixel point in an infrared thermal image is a noise point and an edge point, further carries out weighted mean value denoising according to the probability of the noise point, improves the denoising effect while protecting the edge of the image, and finally analyzes the high-quality infrared thermal image to judge whether the equipment normally operates. Therefore, the accuracy of the equipment abnormity detection of the numerical control ingot splitting machine can be improved.
The invention aims to provide an equipment abnormity detection method for a numerical control ingot splitting machine, which comprises the following steps:
acquiring a gray scale image of an infrared thermal image of the numerical control ingot slicing machine;
traversing the gray-scale map by using a window; classifying the pixel points in the window according to the mean value of the gray value of the pixel points in each window;
acquiring a first probability that the central pixel point in each window is a noise pixel point according to the gray values of other pixel points in the same category as the central pixel point in each window;
the sum of the gray gradient vectors of each pixel point in different directions is used as the heat conduction guiding quantity of each pixel point; according to the heat conduction guiding quantity of each pixel point in each window, which is of the same type as the central pixel point, acquiring a second probability that the central pixel point in each window is a noise pixel point;
acquiring the comprehensive probability of the central pixel point in each window as a noise pixel point according to the first probability and the second probability of the central pixel point in each window;
acquiring the probability that the central pixel point of each window is an edge pixel point according to the comprehensive probability and the gray value of each pixel point in each window;
acquiring the size of a filtering window of each pixel point according to the probability and the comprehensive probability of each pixel point being an edge pixel point;
carrying out mean value filtering on each pixel point according to the size of a filtering window of each pixel point to obtain a denoised gray scale image;
and judging whether the equipment of the numerical control ingot splitting machine is abnormal or not according to the average value of the gray scale of the denoised gray scale image.
In an embodiment, the first probability that the center pixel in each window is a noise pixel is obtained according to the following steps:
acquiring a gray level sequence with gray levels in the gray level image in sequence from small to large;
marking other pixel points in the category of the central pixel point in each window as central pixel points;
respectively acquiring the distance from each gray level in the gray level sequence to the first gray level if the distance between adjacent gray levels in the gray level image is set as 1;
and acquiring a first probability that the central pixel point in each window is a noise pixel point according to the distance from the corresponding gray level of the central pixel point in each window to the first gray level and the distance from the corresponding gray level of each central pixel point in each window to the first gray level.
In an embodiment, the probability that the center pixel of each window is the edge pixel is obtained according to the following steps:
acquiring a gray threshold value for reclassifying the pixels in each window according to the comprehensive probability and gray value of each pixel in each window as a noise pixel;
dividing the pixels in each window into first-class pixels and second-class pixels according to the gray threshold value of the reclassification of the pixels in each window;
and acquiring the probability that the central pixel point in each window is an edge pixel point according to the comprehensive probability and the gray value of each first-class pixel point in each window and the comprehensive probability and the gray value of each second-class pixel point.
In an embodiment, the acquiring the denoised gray scale map further includes:
weighting the gray value of each pixel point in each filtering window by utilizing the comprehensive probability of each pixel point in each filtering window, and obtaining the gray average value in each filtering window through summation;
and replacing the gray value of the central pixel point in each filtering window by using the gray value average value in each filtering window in sequence to obtain a denoised gray image.
In an embodiment, the size of the filtering window of each pixel point is obtained according to the following steps:
acquiring the selection probability of each pixel point to a filtering window according to the probability and the comprehensive probability of each pixel point being an edge pixel point;
and obtaining the size of the filtering window of each pixel point according to the selection probability of each pixel point to the filtering window.
In an embodiment, the size of the filtering window of each pixel point is obtained according to the following steps:
when the selection probability is less than or equal to
The size of the filtering window is 3 multiplied by 3 for the corresponding pixel points;
when the selection probability is greater than
And is smaller than
The size of the filtering window is 5 multiplied by 5 if the corresponding pixel point is selected;
when the selection probability is greater than or equal to
And the size of the filtering window is 7 multiplied by 7 for the corresponding pixel point.
In one embodiment, whether the device is abnormally operated is determined according to the following steps:
acquiring a gray level image of an infrared thermal image of the numerical control ingot slicing machine at certain intervals;
acquiring an absolute value of a difference value of the gray level mean values of the denoised gray level images corresponding to the two adjacent gray level images;
and when the absolute value of the difference value of the gray level mean values is larger than a preset threshold value, judging that the equipment of the numerical control ingot slicing machine is abnormal in operation.
In an embodiment, the second probability that the center pixel point in each window is the noise pixel point is obtained according to the following steps:
acquiring the sum of heat conduction vectors of other pixel points of the same type as the central pixel point in each window;
and acquiring a second probability that the central pixel point in each window is a noise pixel point according to an included angle formed by adding the heat conduction guide quantity of the central pixel point in each window and the heat conduction vectors of other pixel points of the same type of the central pixel point.
The invention has the beneficial effects that:
the invention provides a method for detecting equipment abnormity of a numerical control ingot slicing machine, which is characterized in that an infrared thermal image of the running numerical control ingot slicing machine is collected, self-adaptive selection of a window is realized according to the probability that each pixel point in the image is a noise point and an edge point, weighted mean value denoising is further carried out according to the probability of the noise point, a high-quality infrared thermal image is obtained, and finally, whether the equipment normally runs or not is judged through data analysis of the image, so that the condition that the reliability of an equipment abnormity detection result cannot be ensured due to direct analysis of the collected low-quality infrared thermal image is avoided; therefore, the accuracy of the equipment abnormity detection of the numerical control ingot splitting machine is realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart showing the general steps of an embodiment of an equipment abnormality detection method for a numerically controlled ingot slicing machine according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The method mainly aims at low contrast and large influence of noise on the infrared thermal image, so that the collected low-quality infrared thermal image is directly analyzed, and the reliability of an equipment abnormality detection result cannot be ensured. Therefore, the method realizes the self-adaptive selection of the window by processing the acquired running infrared thermal image of the numerical control ingot slicing machine according to the probability that each pixel point in the image is a noise point and an edge point, and further carries out weighted mean denoising according to the probability of the noise point to obtain the high-quality infrared thermal image. And finally, judging whether the equipment normally operates or not by analyzing the data of the image.
According to the invention, because the numerical control ingot slicing machine needs to carry out long-time cutting operation, and the heat production and heat dissipation processes of the machine body are usually kept in dynamic balance when the machine body runs normally, the temperature change of the equipment is monitored in real time through the infrared thermal image, and whether the inside of the equipment in running is abnormal or not is detected.
The method uses the improved mean filtering to carry out denoising processing on the image, needs a larger window for noise pixel points, reduces the number proportion of the noise pixel points in the window, and improves the denoising effect, and needs a smaller window for edge pixel points, increases the number proportion of the edge pixel points in the window, and improves the edge protection effect.
The invention provides an equipment abnormity detection method for a numerical control ingot splitting machine, which is shown in figure 1 and comprises the following steps:
s1, acquiring a gray scale image of an infrared thermal image of a numerical control ingot slicing machine;
in the embodiment, the infrared thermal image of the numerical control ingot slicing machine is collected in real time in operation by fixedly mounting an infrared thermal imager near the numerical control ingot slicing machine. Because the infrared thermal image is greatly influenced by noise, the quality of the image is greatly reduced, so that the collected infrared thermal image is directly analyzed, and the reliability of the equipment abnormality detection result cannot be ensured. The infrared thermal image needs to be denoised to obtain a high-quality image. To obtain high quality images, the infrared thermal images are first grayed outAnd (5) carrying out degree treatment. And then counting the gray level histogram of the acquired infrared thermal image of the numerical control ingot slicing machine. Obtaining a sequence of pixel gray levels within a gray histogram from small to large
Where n represents the number of gray levels in the gray histogram.
S2, acquiring a first probability that a center pixel point in each window is a noise pixel point;
traversing the gray-scale map by using a window; classifying the pixels in each window according to the mean value of the gray values of the pixels in each window;
acquiring a first probability that the central pixel point in each window is a noise pixel point according to the gray values of other pixel points in the same category as the central pixel point in each window;
the first probability that the central pixel point in each window is the noise pixel point is obtained according to the following steps:
acquiring a gray level sequence with gray levels in the gray level image in a sequence from small to large;
marking other pixel points in the category of the central pixel point in each window as central pixel points;
respectively acquiring the distance from each gray level to the first gray level in the gray level sequence if the distance between adjacent gray levels in the gray level image is set to be 1;
and acquiring a first probability that the central pixel point in each window is a noise pixel point according to the distance from the corresponding gray level of the central pixel point in each window to the first gray level and the distance from the corresponding gray level of each central pixel point in each window to the first gray level.
In this embodiment, the probability that each pixel point in the infrared thermal image of the numerical control ingot slicing machine is a noise pixel point is calculated, a 5 × 5 window is used to traverse the image from top to bottom and from left to right, and the probability that the central pixel point is the noise pixel point is judged according to the similarity of the gray values of the pixel points in the window. The size of the 5 × 5 window is selected to prevent the amount of data in the window from being too small or too large to affect the accuracy of the subsequent analysis.
The known infrared thermal image represents the temperature change of an object, so when a window is positioned in an intersection area with a larger temperature difference in the infrared thermal image, pixel points at two temperatures exist in the window, the gray value difference of the pixel points is larger, the judgment that the central pixel points of the window are noise pixel points on the similarity of the gray values of the pixel points in the window is influenced, therefore, the pixel points in the window need to be classified firstly, the similar pixel points are analyzed, and the classification mode is as follows:
calculating the average gray value of the pixels in the window as B, dividing the pixels in the window into two types of pixels with the gray value less than or equal to B and the gray value greater than B by taking the value of B as a threshold value, taking the pixel of the type of the center pixel of the window as a center type pixel, and marking the center type pixel as the center type pixel
(ii) a Regardless of whether the windows are located in the same temperature region or at an intersection of different temperature regions within the image,
the similar pixel points are similar pixel points at the same temperature.
Because the heat is conducted from high temperature to low temperature, the higher the temperature is, the larger the gray level of the pixel point in the infrared thermal image is, the temperature gradually rises from left to right on the gray level histogram, that is, the gray levels of the pixel points in the region with similar temperature in the image are relatively gathered on the gray level histogram.
Order-of-view pixel gray scale sequence
Respectively acquiring the distance from each gray level to the first gray level if the distance between the middle adjacent gray levels is 1; the distance represented by the gray scale of each pixel from small to large is
. If the central pixel point of the window is a normal point, the window is divided at the same temperature
The distance between the pixel gray levels of the similar pixel points is close, and if the central pixel point of the window is a noise point, the window is divided under the same temperature
The distance between the central pixel point and the pixel gray level of the non-central pixel point in the similar pixel point is far, and the distance is not influenced by the gray value of the pixel point. Therefore, the central pixel point of each window is the first probability of the noise pixel point in the gray scale similarity
The calculation formula is as follows:
in the formula (I), the compound is shown in the specification,
indicating the distance from the gray level corresponding to the y-th pixel point in the infrared thermal image to the first gray level,
indicating that the y-th pixel is centred
In the window
The distance from the corresponding gray level of the z-th non-central pixel point in the class pixel points to the first gray level is obtained, and the difference between the two indicates the distance between the central pixel point and the non-central pixel point in the window on the gray level histogram; q represents within the window
The number of non-center pixels among the class pixels; n-1 represents the distance between the maximum and minimum gray levels, representing grayA maximum distance on the degree histogram, where n represents the number of gray levels within the gray level histogram; order to
Has a value range of [0,1 ]]And m represents the number of pixel points in the infrared thermal image.
Expressing a first probability that a central pixel point y in the window is a noise pixel point in gray scale similarity; therefore, it is
The larger the value of (d), the larger the first probability that the corresponding pixel point is a noise pixel point in the gray scale similarity.
S3, acquiring a second probability that the central pixel point in each window is a noise pixel point;
it should be noted that the conduction of heat is directional, so that the window is within
The heat conduction directions of the pixel-like points are similar, and the noise points can destroy the heat conduction directions in the image. Therefore, the sum of the gray gradient vectors of each pixel point in different directions is used as the heat conduction guiding quantity of each pixel point; according to the heat conduction guidance quantity of each pixel point in the same type as the center pixel point in each window, acquiring a second probability that the center pixel point in each window is a noise pixel point; the second probability that the central pixel point in each window is a noise pixel point in the heat conduction direction is obtained;
specifically, the second probability that the central pixel point in each window is the noise pixel point is obtained according to the following steps:
acquiring the sum of heat conduction vectors of other pixel points of the same type as the central pixel point in each window;
and acquiring a second probability that the central pixel point in each window is a noise pixel point according to an included angle formed by adding the heat conduction guide quantity of the central pixel point in each window and the heat conduction vectors of other pixel points of the same type of the central pixel point.
In this embodiment, the gray gradients from each pixel point in the infrared thermal image to the eight neighborhood directions thereof are counted, and the sum of the gradient vectors is calculated as
The heat conduction vector of each pixel point is expressed by the vector. Therefore, the second probability that the central pixel point in each window is the noise pixel point is known
The calculation formula of (c) is as follows:
in the formula (I), the compound is shown in the specification,
representing the heat conduction vector of the y-th pixel point in the infrared thermal image as the central pixel point in the window,
indicating the inside of a window centered on the y-th pixel
The sum of the heat conduction vectors of the non-center ones of the class pixels,
indicating window central pixel point heat transfer guide quantity and window interior
The included angle between the heat conduction guide quantity and the non-central pixel point in the similar pixel points is smaller, and the smaller the included angle is, the more the description window is
The more similar the heat conduction directions of the pixel-like points are, the smaller the probability of the pixel point being noise is, and 180 represents the directionMeasure the maximum included angle
Has a value range of [0,1 ]]And m represents the number of pixel points in the infrared thermal image.
The second probability that the center pixel y in the window is a noise pixel in the heat conduction direction is represented, so
The larger the value of (a), the larger the second probability that the corresponding pixel point is a noise pixel point in the heat conduction direction.
S4, acquiring the comprehensive probability that each pixel point is a noise pixel point;
acquiring the comprehensive probability of the central pixel point in each window as a noise pixel point according to the first probability and the second probability of the central pixel point in each window; sequentially acquiring the comprehensive probability that each pixel point is a noise pixel point;
in this embodiment, the integrated probability that each pixel in the infrared thermal image is a noise pixel
The calculation formula is as follows:
in the formula (I), the compound is shown in the specification,
expressing a first probability that a central pixel point y in the window is a noise pixel point in gray scale similarity;
representing a second probability that the central pixel point y in the window is a noise pixel point in the heat conduction direction; both values are [0,1 ]]M represents the number of pixel points in the infrared thermal image;
the comprehensive probability that a central pixel point y in the window is a noise pixel point is expressed, and the probability that each pixel point is noise is comprehensively considered by correspondingly giving the same weight of 0.5 to each pixel point in the gray level similarity and the noise probability in the conduction direction by utilizing the probability that each pixel point is noise; therefore, it is
The larger the pixel point is, the larger the probability of the pixel point being noise is, and the value range of the pixel point is [0,1 ]]。
S5, obtaining the probability that each pixel point is an edge pixel point;
acquiring the probability that the central pixel point of each window is an edge pixel point according to the comprehensive probability and the gray value of each pixel point in each window; sequentially acquiring the probability that each pixel point is an edge pixel point;
the probability that the center pixel point of each window is the edge pixel point is obtained according to the following steps:
acquiring a gray threshold value for reclassifying the pixels in each window according to the comprehensive probability and gray value of each pixel in each window as a noise pixel;
dividing the pixels in each window into first-class pixels and second-class pixels according to the gray threshold value of the reclassification of the pixels in each window;
and acquiring the probability that the central pixel point in each window is an edge pixel point according to the comprehensive probability and the gray value of each first-class pixel point in each window and the comprehensive probability and the gray value of each second-class pixel point.
It should be noted that, pixels at two temperatures exist in the window of the edge pixel, and the difference between the gray values of the two types of pixels is large. Due to the influence of noise, the reliability of the gray value of the pixel point is reduced, and certain errors exist in the initial classification of the pixel points in the window, so that the pixel points in the window need to be reclassified. Gray threshold for pixel classification in each 5 x 5 window
The calculation formula is as follows:
in the formula (I), the compound is shown in the specification,
representing the probability that the xth pixel point in a 5 x 5 window centered on the yth pixel point in the infrared thermal image is a noise pixel point, t representing the number of pixel points in the 5 x 5 window centered on the yth pixel point,
expressing the gray value of the xth pixel point in a 5 multiplied by 5 window taking the yth pixel point as the center; m represents the number of pixel points in the infrared thermal image;
representing the weight of each pixel point; when the probability that a pixel is a noise pixel is larger, the reliability of the gray value of the pixel is lower, so that the weight of the gray value of the pixel is smaller, and the weighted average value of the gray value is smaller
And removing the gray threshold value under the influence of the noise in the corresponding window.
Therefore, the pixel points in each window are subjected to corresponding gray threshold values
Are divided into two categories, one is that the gray value is less than
Is marked as a first type of pixel
The other is that the gray value is greater than or equal to
The second kind of pixel points are marked as
. Therefore, the probability that the center pixel point of each window is the edge pixel point
The calculation formula is as follows:
in the formula (I), the compound is shown in the specification,
and
respectively represent the inside of a 5 multiplied by 5 window which takes the y-th pixel point as the center in the infrared thermal image
And
the first of the class pixel points
And a first step of
Probability that each pixel is a noise pixel;
and
respectively represent the inside of a window with the y-th pixel point as the center
And
class pixel point number
And a first
The gray value of each pixel point;
and
respectively represent the inside of a window with the y-th pixel point as the center
And
the number of class pixels; therefore, it is
Within a presentation window
The noise probability gray-scale weighted mean of the class pixels,
within a presentation window
Noise probability gray scale weighted mean value of the class pixel points; the larger the difference between the two mean values is, the larger the probability that two temperature pixel points exist in the window is, namely the larger the probability that the center pixel point of the window is the edge pixel point is. The difference between the two averages is larger than the sum of the two averages, so that the influence of the gray scale difference of the high-temperature region edge larger than that of the low-temperature region edge can be reduced, and the gray scale difference of the low-temperature region edge can be reduced
Has a value range of [0,1 ]]；
The probability that the central pixel y in the window is the edge pixel is expressed, so
The larger the value of (a), the larger the probability that the corresponding pixel point is an edge pixel point.
So far, obtaining a comprehensive probability set of each pixel point in the infrared thermal image as a noise pixel point and a probability set of each pixel point after removing noise influence as an edge pixel point, wherein the probability sets are respectively
And
wherein m represents the number of pixel points in the infrared thermal image, and
and
has a value range of [0,1 ]]。
S6, obtaining the size of a filtering window of each pixel point according to the probability and the comprehensive probability that each pixel point is an edge pixel point;
the filtering window size of each pixel point is obtained according to the following steps:
acquiring the selection probability of each pixel point to a filtering window according to the probability and the comprehensive probability that each pixel point is an edge pixel point;
and obtaining the size of the filtering window of each pixel point according to the selection probability of each pixel point to the filtering window.
In the present embodiment, the selection probability of each pixel point for the filtering window
The calculation formula is as follows:
in the formula (I), the compound is shown in the specification,
the comprehensive probability that the y-th pixel point in the infrared thermal image is a noise pixel point is represented, and the larger the value of the comprehensive probability is, the larger the window size required by the pixel point is;
expressing the probability that the y-th pixel point is an edge pixel point; the larger the value is, the smaller the window size required by the pixel point is;
represents the selection probability of the y pixel point to the filtering window, therefore
The larger the window size needed by the pixel point is; m represents the number of pixel points in the infrared thermal image.
The size of the filtering window of each pixel point is obtained according to the following steps: when the selection probability is less than or equal to
The size of the filtering window is 3 multiplied by 3 if the corresponding pixel point is adopted; when the selection probability is greater than
And is less than
The size of the filtering window is 5 multiplied by 5 if the corresponding pixel point is selected; when the selection probability is greater than or equal to
Corresponding toAnd if the pixel point is located, the size of the filtering window is 7 multiplied by 7.
In the present embodiment, the types of the filter window sizes set are 3 × 3, 5 × 5, and 7 × 7, respectively, and are selected according to the history data, and the operator can determine the filter window size according to the actual situation. Thus giving
The window of pixel point 3 is given
The window of pixel point 5 is given
7 × 7 windows of pixels.
S7, performing mean value filtering on each pixel point according to the size of a filtering window of each pixel point to obtain a denoised grey-scale image;
in the process of obtaining the denoised gray scale image, the method further comprises the following steps:
weighting the gray value of each pixel point in each filtering window by using the comprehensive probability of each pixel point in each filtering window, and obtaining the gray average value in each filtering window through summation;
and replacing the gray value of the central pixel point in each filtering window by using the gray average value in each filtering window in sequence to obtain a denoised gray image.
In this embodiment, the self-adaptive window size of each pixel point in the infrared thermal image and the probability of the noise pixel point are obtained, and since the probability of the noise pixel point is larger, the reliability of the gray value of the pixel point is lower, the improved mean filtering denoising formula of the infrared thermal image is as follows:
where j represents a window size type,
the corresponding window sizes of the pixels are respectively expressed as 3 × 3, 5 × 5 and 7 × 7.
Representing the probability that the x-th pixel point in the corresponding j-type window with the y-th pixel point as the center in the infrared thermal image is the noise pixel point,
indicating the number of pixels in the corresponding j-type window centered on the y-th pixel,
representing the gray value of the x-th pixel point in the corresponding j-type window with the y-th pixel point as the center,
and representing the weight of each pixel point in a corresponding j-type window with the y-th pixel point as the center, wherein m represents the number of the pixel points in the infrared thermal image. When the probability that the pixel is a noise pixel is higher, the reliability of the gray value of the pixel is lower, so that the weight of the gray value of the pixel is lower;
representing the mean value of the gray levels within each filter window.
Finally, the noise probability weighted gray level mean is used
And replacing the gray value of the original central pixel point in the corresponding window to complete the denoising treatment of the infrared thermal image, and acquiring a denoised gray image, namely the high-quality numerical control ingot splitting machine infrared thermal image.
And S8, judging whether the equipment of the numerical control ingot slicing machine is abnormal or not according to the average value of the gray level of the denoised gray level image.
Whether the equipment is abnormal or not is judged according to the following steps:
acquiring a gray level image of an infrared thermal image of the numerical control ingot slicing machine at certain intervals;
acquiring an absolute value of a difference value of the gray level mean values of the denoised gray level images corresponding to the two adjacent gray level images;
and when the absolute value of the difference value of the gray mean values is larger than a preset threshold value, judging that the equipment of the numerical control ingot splitting machine is abnormal in operation.
In this embodiment, according to the obtained denoised grayscale image, since the position of the numerical control ingot slicing machine device is fixed, and the shooting position and angle of the infrared thermal imager are fixed, the position of the cutting servo motor inside the numerical control ingot slicing machine in the infrared thermal image can be selected through the manual frame, the pixel grayscale mean value in the cutting servo motor region in the infrared thermal image is counted once every 10 minutes, the absolute value of the difference value of the two counted adjacent grayscale mean values when the device normally operates in one day is calculated, the mean value of the group of data is taken as K, 1.5K is taken as a threshold, if the absolute value of the difference value of the pixel grayscale mean value in the cutting servo motor region in the infrared thermal image collected twice every 10 minutes is greater than 1.5K, it is determined that the device operates abnormally, and the shutdown inspection is immediately performed.
The invention provides a method for detecting equipment abnormity of a numerical control ingot slicing machine, which is characterized in that an infrared thermal image of the running numerical control ingot slicing machine is collected, self-adaptive selection of a window is realized according to the probability that each pixel point in the image is a noise point and an edge point, weighted mean value denoising is further carried out according to the probability of the noise point, a high-quality infrared thermal image is obtained, and finally, whether the equipment normally runs or not is judged through data analysis of the image, so that the condition that the reliability of an equipment abnormity detection result cannot be ensured due to direct analysis of the collected low-quality infrared thermal image is avoided; therefore, the accuracy of the equipment abnormity detection of the numerical control ingot splitting machine is realized.
The present invention is not limited to the above preferred embodiments, and any modifications, equivalent substitutions, improvements, etc. within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (8)
1. An equipment abnormity detection method for a numerical control ingot splitting machine is characterized by comprising the following steps:
acquiring a gray scale image of an infrared thermal image of the numerical control ingot slicing machine;
traversing the gray-scale image by using a window; classifying the pixels in each window according to the mean value of the gray values of the pixels in each window;
acquiring a first probability that the central pixel point in each window is a noise pixel point according to the gray values of other pixel points in the same category as the central pixel point in each window;
the sum of the gray gradient vectors of each pixel point in different directions is used as the heat conduction guiding quantity of each pixel point; according to the heat conduction guidance quantity of each pixel point in the same type as the center pixel point in each window, acquiring a second probability that the center pixel point in each window is a noise pixel point;
acquiring the comprehensive probability of the central pixel point in each window as a noise pixel point according to the first probability and the second probability of the central pixel point in each window;
acquiring the probability that the central pixel point of each window is an edge pixel point according to the comprehensive probability and the gray value of each pixel point in each window;
acquiring the size of a filtering window of each pixel point according to the probability and the comprehensive probability of each pixel point being an edge pixel point;
carrying out mean value filtering on each pixel point according to the size of a filtering window of each pixel point to obtain a denoised gray scale image;
and judging whether the equipment of the numerical control ingot splitting machine is abnormal or not according to the average value of the gray scale of the denoised gray scale image.
2. The equipment anomaly detection method for the numerical control spindle splitting machine according to claim 1, characterized in that the first probability that the center pixel point in each window is a noise pixel point is obtained according to the following steps:
acquiring a gray level sequence with gray levels in the gray level image in a sequence from small to large;
marking other pixel points in the category of the central pixel point in each window as central pixel points;
respectively acquiring the distance from each gray level in the gray level sequence to the first gray level if the distance between adjacent gray levels in the gray level image is set as 1;
and acquiring a first probability that the central pixel point in each window is a noise pixel point according to the distance from the corresponding gray level of the central pixel point in each window to the first gray level and the distance from the corresponding gray level of each central pixel point in each window to the first gray level.
3. The method for detecting the equipment abnormality of the numerical control ingot slicing machine according to claim 1, wherein the probability that the center pixel point of each window is the edge pixel point is obtained according to the following steps:
acquiring a gray threshold value of reclassification of the pixel points in each window according to the comprehensive probability and gray value of each pixel point in each window being a noise pixel point;
dividing the pixels in each window into first-class pixels and second-class pixels according to the reclassified gray threshold of the pixels in each window;
and acquiring the probability that the central pixel point in each window is an edge pixel point according to the comprehensive probability and the gray value of each first-class pixel point in each window and the comprehensive probability and the gray value of each second-class pixel point.
4. The method for detecting the equipment abnormity of the numerical control ingot splitting machine according to claim 1, wherein in the process of acquiring the de-noised gray scale image, the method further comprises the following steps:
weighting the gray value of each pixel point in each filtering window by using the comprehensive probability of each pixel point in each filtering window, and obtaining the gray average value in each filtering window through summation;
and replacing the gray value of the central pixel point in each filtering window by using the gray average value in each filtering window in sequence to obtain a denoised gray image.
5. The equipment anomaly detection method for the numerical control ingot slicing machine according to claim 1, wherein the size of the filtering window of each pixel point is obtained according to the following steps:
acquiring the selection probability of each pixel point to a filtering window according to the probability and the comprehensive probability that each pixel point is an edge pixel point;
and obtaining the size of the filtering window of each pixel point according to the selection probability of each pixel point to the filtering window.
6. The equipment abnormality detection method for the numerical control ingot slicing machine according to claim 5, wherein the size of the filter window of each pixel point is obtained by the following steps:
when the selection probability is less than or equal to
The size of the filtering window is 3 multiplied by 3 for the corresponding pixel points;
when the selection probability is greater than
And is smaller than
The size of the filtering window is 5 multiplied by 5 for the corresponding pixel point;
when the selection probability is greater than or equal to
And the size of the filtering window is 7 multiplied by 7 for the corresponding pixel point.
7. The method for detecting the abnormality of the apparatus for the numerically controlled ingot slicer as claimed in claim 1, wherein whether the operation of the apparatus is abnormal is determined according to the following steps:
acquiring a gray scale image of the infrared thermal image of the numerical control ingot slicing machine at certain intervals;
acquiring an absolute value of a difference value of the gray level mean values of the denoised gray level images corresponding to the two adjacent gray level images;
and when the absolute value of the difference value of the gray level mean values is larger than a preset threshold value, judging that the equipment of the numerical control ingot slicing machine is abnormal in operation.
8. The equipment abnormality detection method for the numerical control ingot slicing machine according to claim 1, wherein the second probability that the center pixel point in each window is a noise pixel point is obtained according to the following steps:
acquiring the sum of heat conduction vectors of other pixel points of the same type as the central pixel point in each window;
and acquiring a second probability that the central pixel point in each window is a noise pixel point according to an included angle formed by adding the heat conduction guide quantity of the central pixel point in each window and the heat conduction vectors of other pixel points of the same type of the central pixel point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211365049.8A CN115471503B (en) | 2022-11-03 | 2022-11-03 | Equipment abnormity detection method for numerical control ingot splitting machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211365049.8A CN115471503B (en) | 2022-11-03 | 2022-11-03 | Equipment abnormity detection method for numerical control ingot splitting machine |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115471503A true CN115471503A (en) | 2022-12-13 |
| CN115471503B CN115471503B (en) | 2023-04-07 |
Family
ID=84336385
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211365049.8A Active CN115471503B (en) | 2022-11-03 | 2022-11-03 | Equipment abnormity detection method for numerical control ingot splitting machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115471503B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116012346A (en) * | 2023-02-01 | 2023-04-25 | 江苏天南电力股份有限公司 | Power transmission line monitoring method and system |
| CN116342610A (en) * | 2023-05-31 | 2023-06-27 | 山东恒海钢结构有限公司 | Steel structure assembly type building welding abnormality detection method |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008082470A1 (en) * | 2007-01-05 | 2008-07-10 | Raytheon Company | Target recognition system and method |
| CN105243649A (en) * | 2015-11-09 | 2016-01-13 | 天津大学 | Image denoising method based on secondary noise point detection |
| CN109584174A (en) * | 2019-01-29 | 2019-04-05 | 电子科技大学 | A kind of gradient minimum method infrared image edge holding denoising method |
| CN112101365A (en) * | 2020-09-10 | 2020-12-18 | 国网辽宁省电力有限公司电力科学研究院 | Power equipment key feature extraction method and system based on infrared thermal image processing |
| WO2021217643A1 (en) * | 2020-04-30 | 2021-11-04 | 深圳市大疆创新科技有限公司 | Method and device for infrared image processing, and movable platform |
| CN114220084A (en) * | 2021-11-11 | 2022-03-22 | 国网浙江省电力有限公司嘉兴供电公司 | Distribution equipment defect identification method based on infrared image |
| CN115131351A (en) * | 2022-08-31 | 2022-09-30 | 微山宏捷机械有限公司 | Engine oil radiator detection method based on infrared image |
| CN115170576A (en) * | 2022-09-09 | 2022-10-11 | 山东中发新材料科技有限公司 | Aluminum pipe surface defect detection method based on machine vision |
-
2022
- 2022-11-03 CN CN202211365049.8A patent/CN115471503B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008082470A1 (en) * | 2007-01-05 | 2008-07-10 | Raytheon Company | Target recognition system and method |
| CN105243649A (en) * | 2015-11-09 | 2016-01-13 | 天津大学 | Image denoising method based on secondary noise point detection |
| CN109584174A (en) * | 2019-01-29 | 2019-04-05 | 电子科技大学 | A kind of gradient minimum method infrared image edge holding denoising method |
| WO2021217643A1 (en) * | 2020-04-30 | 2021-11-04 | 深圳市大疆创新科技有限公司 | Method and device for infrared image processing, and movable platform |
| CN112101365A (en) * | 2020-09-10 | 2020-12-18 | 国网辽宁省电力有限公司电力科学研究院 | Power equipment key feature extraction method and system based on infrared thermal image processing |
| CN114220084A (en) * | 2021-11-11 | 2022-03-22 | 国网浙江省电力有限公司嘉兴供电公司 | Distribution equipment defect identification method based on infrared image |
| CN115131351A (en) * | 2022-08-31 | 2022-09-30 | 微山宏捷机械有限公司 | Engine oil radiator detection method based on infrared image |
| CN115170576A (en) * | 2022-09-09 | 2022-10-11 | 山东中发新材料科技有限公司 | Aluminum pipe surface defect detection method based on machine vision |
Non-Patent Citations (4)
| Title |
|---|
| DANG N.H.THANH ET AL: "Adaptive Switching Weight Mean Filter for Salt and Pepper Image Denoising", 《PROCEDIA COMPUTER SCIENCE》 * |
| LANGWU ET AL: "Infrared small target detection based on gray intensity descent and local gradient watershed", 《INFRARED PHYSICS & TECHNOLOGY》 * |
| 孙殿柱等: "散乱点云自适应滤波算法", 《北京航空航天大学学报》 * |
| 罗恒亮等: "中值滤波在红外成像引信中的应用及硬件实现", 《激光与红外》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116012346A (en) * | 2023-02-01 | 2023-04-25 | 江苏天南电力股份有限公司 | Power transmission line monitoring method and system |
| CN116342610A (en) * | 2023-05-31 | 2023-06-27 | 山东恒海钢结构有限公司 | Steel structure assembly type building welding abnormality detection method |
| CN116342610B (en) * | 2023-05-31 | 2023-08-15 | 山东恒海钢结构有限公司 | Steel structure assembly type building welding abnormality detection method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115471503B (en) | 2023-04-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115471503B (en) | Equipment abnormity detection method for numerical control ingot splitting machine | |
| CN114549522A (en) | Textile quality detection method based on target detection | |
| CN113724231B (en) | Industrial defect detection method based on semantic segmentation and target detection fusion model | |
| CN111815564B (en) | Method and device for detecting silk ingots and silk ingot sorting system | |
| CN116485779B (en) | Adaptive wafer defect detection method and device, electronic equipment and storage medium | |
| CN115311267B (en) | Method for detecting abnormity of check fabric | |
| CN116777907A (en) | Sheet metal part quality detection method | |
| CN115063430B (en) | Electric pipeline crack detection method based on image processing | |
| CN115063409A (en) | Method and system for detecting surface material of mechanical cutter | |
| CN116740058B (en) | Quality detection method for solid state disk matched wafer | |
| CN117173187A (en) | Intelligent valve fault detection system | |
| CN115311484B (en) | Optimized clustering segmentation method for weld data | |
| CN116363126B (en) | Welding quality detection method for data line USB plug | |
| CN115980050A (en) | Water quality detection method and device for water outlet, computer equipment and storage medium | |
| CN117078679B (en) | Automatic assembly line production detection method for cooling fan based on machine vision | |
| CN111695582A (en) | Flutter texture detection method and device | |
| CN114202544B (en) | Complex workpiece defect detection method based on self-encoder | |
| US20240005477A1 (en) | Index selection device, information processing device, information processing system, inspection device, inspection system, index selection method, and index selection program | |
| CN112581472B (en) | Target surface defect detection method facing human-computer interaction | |
| CN115049641A (en) | Electric data processing method and system for anomaly detection of mechanical parts | |
| CN114663365A (en) | Defect detection method, device and computer storage medium | |
| CN108181315B (en) | Image processing-based biscuit damage detection device and detection method | |
| CN112150434A (en) | Tire defect detection method, device, equipment and storage medium | |
| CN111681229A (en) | Deep learning model training method, wearable clothes flaw identification method and wearable clothes flaw identification device | |
| CN111724352B (en) | Patch LED flaw labeling method based on kernel density estimation |
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 |