CN117173208B - Error judgment method for sawing process of aluminum profile based on data analysis - Google Patents
Error judgment method for sawing process of aluminum profile based on data analysis Download PDFInfo
- Publication number
- CN117173208B CN117173208B CN202311441203.XA CN202311441203A CN117173208B CN 117173208 B CN117173208 B CN 117173208B CN 202311441203 A CN202311441203 A CN 202311441203A CN 117173208 B CN117173208 B CN 117173208B
- Authority
- CN
- China
- Prior art keywords
- saw blade
- frame
- gray level
- membership
- pixel point
- 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
- 238000000034 method Methods 0.000 title claims abstract description 64
- XAGFODPZIPBFFR-UHFFFAOYSA-N aluminium Chemical compound [Al] XAGFODPZIPBFFR-UHFFFAOYSA-N 0.000 title claims abstract description 37
- 229910052782 aluminium Inorganic materials 0.000 title claims abstract description 37
- 238000007405 data analysis Methods 0.000 title claims abstract description 15
- 230000002159 abnormal effect Effects 0.000 claims abstract description 55
- 238000012216 screening Methods 0.000 claims abstract description 10
- 206010044048 Tooth missing Diseases 0.000 claims abstract description 6
- 230000005856 abnormality Effects 0.000 claims description 25
- 238000006073 displacement reaction Methods 0.000 claims description 4
- 238000001914 filtration Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 abstract description 3
- 239000000203 mixture Substances 0.000 description 6
- 238000005520 cutting process Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000012800 visualization Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000003708 edge detection Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 101100272279 Beauveria bassiana Beas gene Proteins 0.000 description 1
- 238000005299 abrasion Methods 0.000 description 1
- 239000004411 aluminium Substances 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
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
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of image processing, in particular to an aluminum profile sawing process error judging method based on data analysis, which comprises the following steps: collecting continuous frame saw blade gray level images, carrying out mixed Gaussian background modeling on the continuous frame saw blade gray level images, obtaining Gaussian mixed models of each position, obtaining the membership degree of the pixel point of the same position in each frame saw blade gray level image to each sub Gaussian model of the position, constructing a membership sequence of each position, obtaining the abnormal degree of the pixel point of each position in each frame saw blade gray level image according to the membership sequence, screening abnormal pixel points according to the abnormal degree, identifying saw blade tooth missing according to the abnormal pixel points, and carrying out error judgment in the aluminum profile sawing process. The invention eliminates the influence of the motion blur of the saw blade teeth in the sawing process of the aluminum profile, can accurately identify the missing saw blade teeth in the rotating and moving processes of the saw blade, and can timely find errors generated in the sawing process of the aluminum profile.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to an aluminum profile sawing process error judging method based on data analysis.
Background
In the sawing process of the aluminum profile, the saw blade rotates and moves simultaneously to cut the aluminum profile, and if saw blade teeth are worn and lost in the sawing process, the saw blade teeth deform in the size of the sawed aluminum profile, so that cutter errors are caused. It is therefore necessary to identify the missing saw blade teeth during the sawing of the aluminium profile.
Because the saw blade rotates at a high speed and moves in the process, the position of the saw blade teeth in the shot saw blade image generates motion blur, and under the influence of the motion blur, the edge of the saw blade teeth cannot be completely extracted by utilizing methods such as edge detection, threshold segmentation and the like, and further the missing of the saw blade teeth cannot be identified, so that errors generated during sawing of the aluminum profile cannot be timely identified and judged.
Disclosure of Invention
In order to solve the problems, the invention provides an error judging method for an aluminum profile sawing process based on data analysis, which comprises the following steps:
collecting saw blade gray level images of continuous frames;
carrying out mixed Gaussian background modeling on the saw blade gray level images of the continuous frames to obtain Gaussian mixed models of each position in the saw blade gray level images of the continuous frames; acquiring the membership degree of a pixel point at the same position in each frame of saw blade gray level image to each sub-Gaussian model at the position; screening a membership sub-Gaussian model of the pixel point at the same position in each frame of saw blade gray level image according to the membership degree, and acquiring a membership sequence of each position in the saw blade gray level image of the continuous frame according to the membership sub-Gaussian model of the pixel point at the same position in each frame of saw blade gray level image;
obtaining the degree of abnormality of the pixel point at each position in each frame of saw blade gray level image according to the membership sequence, and screening abnormal pixel points according to the degree of abnormality;
and identifying saw blade teeth missing according to the abnormal pixel points, and judging errors in the sawing process of the aluminum profile.
Preferably, the pixel at the same position in each frame of saw blade gray level image is obtainedThe membership degree of the point to each sub-Gaussian model of the position comprises the following specific steps:wherein,representing the membership degree of a pixel point at a j-th position in the i-th frame saw blade gray level image to a k-th sub Gaussian model of the position;a gray value of a pixel point at a j-th position in the i-th frame saw blade gray image;mean parameters of the kth sub-Gaussian model representing the jth position;standard deviation parameters of a kth sub-gaussian model representing a jth position; exp () represents an exponential function based on a natural constant; the absolute value symbol is denoted by i.
Preferably, the filtering the membership sub-gaussian model of the pixel point at the same position in each frame of saw blade gray level image according to the membership degree includes the following specific steps:
and for the pixel point of each position in each frame of saw blade gray level image, acquiring a sub-Gaussian model corresponding to the largest membership degree in the membership degrees of the pixel point to each sub-Gaussian model of the position, and taking the sub-Gaussian model as the membership sub-Gaussian model of the pixel point.
Preferably, the step of obtaining the membership sequence of each position in the saw blade gray level image of the continuous frame according to the membership sub-gaussian model of the pixel point at the same position in each frame of the saw blade gray level image comprises the following specific steps:
and forming a one-dimensional sequence according to the sequence of the membership sub Gaussian model of the pixel point at the same position in all frame saw blade gray images and taking the sequence of the sequence as the membership sequence of the position.
Preferably, said method is based on said membership orderThe column obtains the degree of abnormality of the pixel point at each position in each frame of saw blade gray level image, comprising the following specific steps:in the method, in the process of the invention,representing the degree of abnormality of the pixel point at the j-th position in the i-th frame saw blade gray level image;the number of elements in the membership sequence of the j-th position in the saw blade gray level image of the continuous frame, which is the same as the serial number of the membership sub-Gaussian model of the pixel point of the j-th position in the saw blade gray level image of the i-th frame;the number of elements, before and after the serial number of the membership sub-Gaussian model corresponding to the ith frame of saw blade gray image, which are continuously the same as the serial number in the membership sequence of the jth position in the saw blade gray image of the continuous frame;a serial number of a membership sub-Gaussian model for a pixel point at a j-th position in the i-th frame saw blade gray level image;a serial number of a membership sub-Gaussian model for a pixel point at a j-th position in the (i+1) -th frame saw blade gray level image;a difference function between a membership sub-Gaussian model of a pixel point at a j-th position in the i-th frame of saw blade gray level image and a membership sub-Gaussian model of a pixel point at a j-th position in the i+1-th frame of saw blade gray level image is represented;the pixel point representing the j-th position in the i-th frame saw blade gray level image is relative to the j-th positionMembership of the individual gaussian models;the pixel point at the j-th position in the i-th frame saw blade gray level image corresponds to the j-th positionMembership of the individual gaussian models; m represents a preset local range size.
Preferably, the method for obtaining the difference function is as follows:wherein 1 is the difference between the frame numbers of the ith frame of saw blade gray level image and the (i+1) th frame of saw blade gray level image; exp () represents an exponential function based on a natural constant; the absolute value symbol is denoted by i.
Preferably, the step of identifying missing saw blade teeth according to the abnormal pixel points comprises the following specific steps:
obtaining the displacement x of saw blade teeth after one frame time according to the rotation direction, the rotating speed, the moving direction and the moving speed of the saw blade; projecting abnormal pixel points in each frame of saw blade gray level image into the previous frame of saw blade gray level image, acquiring the distance between each abnormal pixel point and each projection point in the previous frame of saw blade gray level image, and acquiring the distance between the abnormal pixel points and each projection pointA point pair consisting of an abnormal pixel point and a projection point in the range as an abnormal point pair, whereinIs a preset error threshold;
the missing saw blade teeth are identified based on the outlier pairs.
Preferably, the identifying the missing saw blade teeth according to the abnormal point pairs comprises the following specific steps:
and presetting a quantity threshold, and when the quantity of the abnormal point pairs is larger than the quantity threshold, missing saw blade teeth.
Preferably, the step of screening the abnormal pixel points according to the abnormality degree includes the following specific steps:
presetting an abnormal threshold value, and taking a pixel point with the degree of abnormality greater than the abnormal threshold value in each frame of saw blade gray level image as an abnormal pixel point.
The technical scheme of the invention has the beneficial effects that: according to the invention, mixed Gaussian background modeling is carried out on the saw blade gray level images of continuous frames, a Gaussian mixed model of each position is obtained, the membership degree of the pixel point of the same position in each frame of saw blade gray level image to each sub Gaussian model of the position is obtained, a membership sequence of each position is constructed according to the membership sub Gaussian model of the pixel point of the same position in each frame of saw blade gray level image, the abnormal degree of the pixel point of each position in each frame of saw blade gray level image is obtained by analyzing the change rule of the membership sub Gaussian model in the membership sequence, the abnormal pixel point is screened according to the abnormal degree, the saw blade tooth missing is identified according to the abnormal pixel point, and the error judgment of the saw cutting process of the aluminum profile is carried out. The invention eliminates the influence of the motion blur of the saw blade teeth in the sawing process of the aluminum profile, can accurately identify the missing saw blade teeth in the rotating and moving processes of the saw blade, and can timely find errors generated in the sawing process of the aluminum profile.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of the steps of the error judging method of the aluminum profile sawing process based on data analysis;
FIG. 2 is a frame of a blade gray scale image;
FIG. 3 is another frame of a blade gray scale image;
FIG. 4 is a still further frame of blade gray scale image;
FIG. 5 is a schematic diagram of a Gaussian mixture model;
FIG. 6 is an anomaly level visualization of a frame of blade gray scale image;
FIG. 7 is an anomaly level visualization of another frame of blade gray level image;
fig. 8 is an anomaly level visualization of yet another frame of blade gray level image.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following detailed description is given below of the error judging method for the sawing process of the aluminum profile based on data analysis according to the invention, which is provided by combining the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following specifically describes a specific scheme of the error judging method of the aluminum profile sawing process based on data analysis provided by the invention with reference to the accompanying drawings.
Referring to fig. 1, a step flow chart of a method for determining errors in a sawing process of an aluminum profile based on data analysis according to an embodiment of the invention is shown, the method includes the following steps:
s001, acquiring saw blade gray level images of continuous frames.
And erecting a camera at the side of the saw blade of the aluminum profile sawing machine, and shooting saw blade images of continuous frames by using the camera in the process of sawing the aluminum profile by the saw blade. The shot saw blade image is an RGB image, and in order to facilitate subsequent processing, the RGB image is subjected to graying operation, and the RGB image is converted into a gray image and recorded as a saw blade gray image. The multi-frame saw blade gray scale image of this embodiment is shown in fig. 2,3 and 4. In order to ensure that the missing blade teeth can be recognized later, it is necessary to take a high frame rate image when taking a continuous frame of blade images, and the present embodiment is described with a frame rate of 240 frames/second, and is not particularly limited.
To this end, successive frames of blade gray scale images are acquired.
S002, carrying out mixed Gaussian background modeling on the saw blade gray level images of the continuous frames, obtaining a Gaussian mixed model of each position in the saw blade gray level images of the continuous frames, obtaining the membership degree of the pixel point of the same position in each frame of the saw blade gray level images to each sub-Gaussian model of the position, screening the membership sub-Gaussian model of the pixel point of the same position in each frame of the saw blade gray level images according to the membership degree, and obtaining the membership sequence of each position in the saw blade gray level images of the continuous frames according to the membership sub-Gaussian model of the pixel point of the same position in each frame of the saw blade gray level images.
It should be noted that, when the saw blade is in the process of rotating and moving at high speed when the image of the saw blade is shot, when the camera is exposed, the saw blade teeth on the saw blade are blurred due to movement, so that the saw blade teeth in the gray image of the saw blade are unclear, when the camera is exposed, the gray of the saw blade teeth is overlapped with the gray of the background, as in fig. 2,3 and 4, the positions of the edge of the saw blade in the saw blade cannot clearly see the saw blade teeth, but the background can be seen through the edge of the saw blade. Under the influence of motion blur, the edge of the saw blade teeth cannot be completely extracted by utilizing methods such as edge detection, threshold segmentation and the like, and further the missing of the saw blade teeth cannot be identified, so that errors generated during sawing of the aluminum profile cannot be timely identified and judged.
It should be further noted that during rotation and movement of the saw blade, the same position in the blade gray images of successive frames may be different image features in different frames, for example, when the saw blade is not moved to the position, the position in the blade gray image of the corresponding frame is the background, when the saw blade teeth are moved to the position, the position in the blade gray image of the corresponding frame is the superimposed effect of the saw blade teeth and the background, and when the interior of the saw blade is moved to the position, the position in the blade gray image of the corresponding frame is the saw blade. At the moment, mixed Gaussian background modeling can be carried out on the saw blade gray level images of the continuous frames, and the image characteristic change condition corresponding to each position in different frames is analyzed according to the Gaussian mixed model of each position in the saw blade gray level images of the continuous frames, so that abnormal pixel points possibly with saw blade tooth missing can be identified according to the change condition.
In this embodiment, a mixed gaussian background modeling is performed on the blade gray level images of the continuous frames, and a gaussian mixed model of each position in the blade gray level images of the continuous frames is obtained. It should be noted that, the gaussian mixture background modeling is to construct a gaussian mixture model for pixels at the same position in a continuous video frame, and is specifically implemented as a known technology, which is not described in detail in the embodiment of the present invention. Fig. 5 is a schematic diagram of a gaussian mixture model of the same position pixel point in a blade gray scale image of successive frames.
It should be noted that the gaussian mixture model at each position reflects the gray scale distribution of the pixel point at that position at different frames (i.e., at different moments). Each sub-gaussian model corresponding to the gaussian mixture model represents an image feature at that location, such as a background, a saw blade, a superposition of blade teeth and background, etc. In order to acquire the change condition of the image features of the position in different frames, abnormal pixel points are identified according to the change condition, and the membership degree of the pixel points corresponding to the position in each frame of saw blade gray level image to each sub-Gaussian model is required to be acquired.
In this embodiment, the membership degree of the pixel point at the same position in each frame of saw blade gray level image to each sub-gaussian model at the position is obtained:wherein,representing the membership degree of a pixel point at a j-th position in the i-th frame saw blade gray level image to a k-th sub Gaussian model of the position;a gray value of a pixel point at a j-th position in the i-th frame saw blade gray image;mean parameters of the kth sub-Gaussian model representing the jth position;standard deviation parameters of a kth sub-gaussian model representing a jth position; exp () represents an exponential function based on a natural constant; the absolute value symbol; when the standard deviation parameter of the kth sub-Gaussian model at the jth position is larger, the gray scale fluctuation range of the image feature corresponding to the kth sub-Gaussian model at the jth position is larger, and at the moment, if the gray scale value of the pixel point at the jth position in the ith frame of saw blade gray image is closer to the mean value parameter of the kth sub-Gaussian model at the jth position, the pixel point at the jth position in the ith frame of saw blade gray image is more likely to belong to the image feature corresponding to the kth sub-Gaussian model at the jth position, and at the moment, the membership degree of the pixel point at the jth position in the ith frame of saw blade gray image to the kth sub-Gaussian model at the jth position is higher.
And acquiring a sub-Gaussian model corresponding to the largest membership degree in the membership degrees of the pixel point to each sub-Gaussian model of the position as a membership sub-Gaussian model of the pixel point for the pixel point at the j position in the i-th frame saw blade gray level image.
And similarly, obtaining a membership sub-Gaussian model of the pixel point at the j-th position in each frame of saw blade gray level image. And forming a one-dimensional sequence according to the sequence of the membership sub-Gaussian model of the pixel point at the j-th position in all frame saw blade gray images and taking the sequence as the membership sequence of the j-th position.
And similarly, obtaining a membership sequence of each position in the saw blade gray level image of the continuous frames.
S003, obtaining the degree of abnormality of the pixel point at each position in each frame of saw blade gray level image according to the membership sequence, and screening abnormal pixel points according to the degree of abnormality.
It should be noted that, during rotation and movement of the saw blade, if there is no missing saw blade tooth, the corresponding image features in the gray images of the saw blade in the same position are continuous in successive frames, for example, the image features of the position are all background during a period of time before the saw blade moves to the position, the image features of the position are all superposition effects of the saw blade tooth and the background during a period of time after the saw blade tooth moves to the position, and the image features of the position are all saw blade during a period of time after the saw blade moves to the position. In the membership sequence represented in this position, the sequence numbers of the membership sub-Gaussian models are consecutively identical at the local position, e.g., {3,3,3,3,3,3,1,1,1,1,2,2,2,2,2,2,3,3,3}.
When the missing blade tooth rotates to that position, the image feature of that position is background. Because the missing saw blade teeth are missing due to abrasion or tipping of individual teeth, the time of the missing saw blade teeth at the position is very short in the process of rotating the saw blade teeth for one circle, and the image characteristic of the position is the superposition effect of the saw blade teeth and the background in a period of time after the saw blade teeth move to the position, and the image characteristic of the position is the superposition effect of the saw blade teeth and the background in a very small part of time is the background. In the membership sequence represented in this location, the sequence number of the membership sub-Gaussian model jumps at a local location, e.g., {3,3,3,3,3,3,1,1,3,1,2,2,2,2,2,2,3,3,3}. Therefore, the sequence number change condition of the membership sub-Gaussian model in the membership sequence of each position can be analyzed, and the pixel positions possibly with saw blade tooth missing are screened.
In this embodiment, the local range size m is preset to determine local continuity of the membership sub-gaussian model sequence number of the current position in the membership sequence, and the implementation personnel can set the local range size m according to practical implementation conditions, for example, m=10.
For each position in the saw blade gray level images of the continuous frames, constructing a difference function of a membership sub-Gaussian model of the position in the saw blade gray level images of different frames according to a membership sequence of the position:wherein,a difference function between a membership sub-Gaussian model of a pixel point at a j-th position in the i-th frame of saw blade gray level image and a membership sub-Gaussian model of a pixel point at a j-th position in the i+l-th frame of saw blade gray level image is represented;a serial number of a membership sub-Gaussian model for a pixel point at a j-th position in the i-th frame saw blade gray level image;a serial number of a membership sub-Gaussian model for a pixel point at a j-th position in the i+l frame saw blade gray level image; l is the difference between the frame numbers of the i frame saw blade gray level image and the i+l frame saw blade gray level image; exp () represents an exponential function based on a natural constant; the absolute value symbol; when the membership sub-Gaussian model of the pixel point at the j-th position in the i-th frame of saw blade gray level image is the same as the membership sub-Gaussian model of the pixel point at the j-th position in the i+l-th frame of saw blade gray level image,=0; when the membership sub-Gaussian model of the pixel point at the jth position in the ith frame of saw blade gray image is different from the membership sub-Gaussian model of the pixel point at the jth position in the (i+l) th frame of saw blade gray image, the difference degree of the membership sub-Gaussian models is obtained according to the difference between the frame numbers of the ith frame of saw blade gray image and the (i+l) th frame of saw blade gray image, when the difference between the frame numbers is smaller, whether the membership sub-Gaussian models of the pixel point at the jth position in the two frames are the same or not is paid more attention to when the difference between the frame numbers is smallerThe larger and closer to 1. When the difference between the frames is larger, the less attention is paid to whether the membership sub-Gaussian models of the pixel points at the j-th position in the two frames are the same or notThe smaller.
In this embodiment, for each position in the blade gray level image of the continuous frame, the degree of abnormality of the corresponding pixel point of the position in the blade gray level image of each frame is obtained according to the membership sequence of the position:in the method, in the process of the invention,representing the degree of abnormality of the pixel point at the j-th position in the i-th frame saw blade gray level image;the number of elements in the membership sequence of the j-th position in the saw blade gray level image of the continuous frame, which is the same as the serial number of the membership sub-Gaussian model of the pixel point of the j-th position in the saw blade gray level image of the i-th frame;the number of elements which are the same with the sequence number before and after the sequence number of the membership sub-Gaussian model corresponding to the ith frame of saw blade gray image (i.e. the ith element in the membership sequence) in the membership sequence of the jth position in the saw blade gray image of the continuous frame is represented; when (when)When the number of the membership sub-Gaussian model of the pixel point at the j-th position in the i-th frame of saw blade gray level image is larger, the probability that the pixel point at the j-th position in the i-th frame of saw blade gray level image is missing saw blade teeth is smaller, and at the moment, the degree of abnormality is smaller. Conversely, whenWhen the number of the membership sub-Gaussian model of the pixel point at the j position in the i-th frame of saw blade gray level image is smaller, jump occurs in the membership sequence of the j position, and the probability that the pixel point at the j position in the i-th frame of saw blade gray level image is missing saw blade teeth is larger, and the degree of abnormality is larger;
in the method, in the process of the invention,a serial number of a membership sub-Gaussian model for a pixel point at a j-th position in the i-th frame saw blade gray level image;a serial number of a membership sub-Gaussian model for a pixel point at a j-th position in the i+l frame saw blade gray level image;the difference function between the membership sub-Gaussian model of the pixel point at the j-th position in the i-th frame of saw blade gray level image and the membership sub-Gaussian model of the pixel point at the j-th position in the i+l-th frame of saw blade gray level image is represented and used for measuring the consistency of the membership sub-Gaussian model of the pixel point at the j-th position in the i-th frame of saw blade gray level image and the membership sub-Gaussian model of the pixel point at the j-th position in the i+l-th frame of saw blade gray level image; m represents the local range size.
In the method, in the process of the invention,the pixel point representing the j-th position in the i-th frame saw blade gray level image is relative to the j-th positionThe membership degree of the sub-Gaussian model, namely the membership degree of the pixel point at the j-th position in the i-th frame saw blade gray level image to the sub-Gaussian model;the pixel point at the j-th position in the i-th frame saw blade gray level image corresponds to the j-th positionMembership of the sub-Gaussian model, namely membership of a pixel point at a j-th position in an i-th frame saw blade gray level image to a pixel point at a j-th position in an i+l-th frame saw blade gray level image; when (when)Andthe smaller the difference is, the pixel point at the j-th position in the i-th frame saw blade gray level image is corresponding to the j-th positionThe membership degree of the sub-Gaussian model is also large, and the pixel point at the j-th position in the i-th frame saw blade gray level image is more likely to be affected by noiseThe membership sub-Gaussian model is inconsistent with the membership sub-Gaussian model of the pixel point at the j-th position in the (i+l) -th frame saw blade gray level imageNot trusted; when (when)Andwhen the difference is larger, the pixel point at the j-th position in the i-th frame saw blade gray level image is less likely to be affected by noise, and the membership sub-Gaussian model is more likely to be determined by the actually corresponding image characteristics, at the momentTrusted; thus will beAs a means ofWeight of (2), pairAnd correcting so that the finally obtained degree of abnormality is more accurate.
And similarly, acquiring the degree of abnormality of the pixel point at each position in each frame of saw blade gray level image. Fig. 6, 7 and 8 are images of the degree of abnormality of the pixel point at each position in the blade gray scale image corresponding to fig. 2,3 and 4, respectively.
An abnormal threshold T is preset, which is used for screening the abnormal degree, obtaining abnormal pixel points, and an implementation person can set the abnormal threshold T according to the actual implementation situation, for example, t=0.7.
And taking the pixel point with the abnormality degree larger than the abnormality threshold value in each frame of saw blade gray level image as an abnormal pixel point.
Thus, an abnormal pixel point is obtained.
S004, identifying saw blade tooth missing according to the abnormal pixel points, and judging errors in the sawing process of the aluminum profile.
The abnormal pixel points may be an abnormality caused by noise or an abnormality caused by missing saw blade teeth. During rotation and translation of the blade, anomalies due to missing blade teeth also rotate and translate in successive frames of blade gray images, while anomalies due to noise typically only appear in isolation and not in successive frames of blade gray images. Therefore, the distribution position relation of abnormal pixels in the gray level images of the saw blade in continuous frames can be analyzed, and the missing of the saw blade teeth can be identified.
The camera takes the saw blade images of successive frames at a frame rate of 240 frames per second, the time difference between adjacent frames isSecond. The rotation direction, rotation speed, moving direction and moving speed of the saw blade are obtained, and the data can be obtained in parameters of the aluminum profile sawing machine. Obtaining the saw blade tooth passing by according to the rotation direction, the rotation speed, the movement direction and the movement speedThe displacement in seconds is denoted as x. It should be noted that, obtaining the displacement of the saw blade tooth after a certain time according to the rotation direction, the rotation speed, the movement direction and the movement speed is a physical known technique, and will not be described in detail in this embodiment.
Projecting abnormal pixel points in each frame of saw blade gray level image into the previous frame of saw blade gray level image, acquiring the distance between each abnormal pixel point and each projection point in the previous frame of saw blade gray level image, and acquiring the distance between each abnormal pixel point and each projection pointA point pair consisting of an abnormal pixel point and a projection point in the range as an abnormal point pair, whereinFor the preset error threshold, the operator can set the error threshold according to the actual implementation conditionFor example=1。
All abnormal point pairs in the blade gray scale image of successive frames are acquired.
Preset number thresholdFor judging whether the saw blade teeth are missing or not, the operator can set a quantity threshold according to the actual implementation conditionFor example=10. When the number of abnormal point pairs is greater than the number thresholdWhen the saw blade teeth are in missing state, abnormal points are continuously generated in saw blade gray images of different frames, saw cutting errors are generated when saw cutting of the aluminum profile is affected by the saw blade teeth in missing state, saw cutting of the aluminum profile is stopped immediately at the moment, saw blades are replaced, and follow-up saw cutting errors of the aluminum profile are prevented.
Through the steps, the error judgment of the sawing process of the aluminum profile is realized.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (6)
1. The error judging method for the sawing process of the aluminum profile based on data analysis is characterized by comprising the following steps of:
collecting saw blade gray level images of continuous frames;
carrying out mixed Gaussian background modeling on the saw blade gray level images of the continuous frames to obtain Gaussian mixed models of each position in the saw blade gray level images of the continuous frames; acquiring the membership degree of a pixel point at the same position in each frame of saw blade gray level image to each sub-Gaussian model at the position; screening a membership sub-Gaussian model of the pixel point at the same position in each frame of saw blade gray level image according to the membership degree, and acquiring a membership sequence of each position in the saw blade gray level image of the continuous frame according to the membership sub-Gaussian model of the pixel point at the same position in each frame of saw blade gray level image;
obtaining the degree of abnormality of the pixel point at each position in each frame of saw blade gray level image according to the membership sequence, and screening abnormal pixel points according to the degree of abnormality;
identifying saw blade teeth missing according to the abnormal pixel points, and judging errors in the sawing process of the aluminum profile;
the step of obtaining the membership degree of the pixel point at the same position in each frame of saw blade gray level image to each sub-Gaussian model at the position comprises the following specific steps:
wherein d i,j,k Representing the membership degree of a pixel point at a j-th position in the i-th frame saw blade gray level image to a k-th sub Gaussian model of the position; h is a i,j A gray value of a pixel point at a j-th position in the i-th frame saw blade gray image; mu (mu) j,k Mean parameters of the kth sub-Gaussian model representing the jth position; sigma (sigma) j,k Standard deviation parameters of a kth sub-gaussian model representing a jth position; exp () represents an exponential function based on a natural constant; the absolute value symbol;
the method for acquiring the degree of abnormality of the pixel point at each position in each frame of saw blade gray level image according to the membership sequence comprises the following specific steps:
wherein, c i,j Representing the j-th position in the i-th frame blade gray scale imageDegree of abnormality of the pixel points; n (N) i,j The number of elements in the membership sequence of the j-th position in the saw blade gray level image of the continuous frame, which is the same as the serial number of the membership sub-Gaussian model of the pixel point of the j-th position in the saw blade gray level image of the i-th frame; n is n i,j The number of elements, before and after the serial number of the membership sub-Gaussian model corresponding to the ith frame of saw blade gray image, which are continuously the same as the serial number in the membership sequence of the jth position in the saw blade gray image of the continuous frame; s is S i,j A serial number of a membership sub-Gaussian model for a pixel point at a j-th position in the i-th frame saw blade gray level image; s is S i+l,j A serial number of a membership sub-Gaussian model for a pixel point at a j-th position in the i+l frame saw blade gray level image; f (S) i,j ,S i+l,j ) A difference function between a membership sub-Gaussian model of a pixel point at a j-th position in the i-th frame of saw blade gray level image and a membership sub-Gaussian model of a pixel point at a j-th position in the i+l-th frame of saw blade gray level image is represented;s of pixel point representing jth position in ith frame saw blade gray level image for the position i,j Membership of the individual gaussian models; />The pixel point at the j-th position in the i-th frame saw blade gray level image is corresponding to the S-th position of the j-th position i+l,j Membership of the individual gaussian models; m represents a preset local range size;
the method for identifying the saw blade tooth missing according to the abnormal pixel points comprises the following specific steps:
obtaining the displacement x of saw blade teeth after one frame time according to the rotation direction, the rotating speed, the moving direction and the moving speed of the saw blade; projecting abnormal pixel points in each frame of saw blade gray level image into the previous frame of saw blade gray level image, obtaining the distance between each abnormal pixel point in the previous frame of saw blade gray level image and each projection point, and obtaining a point pair formed by the abnormal pixel points and the projection points, wherein the distance is in the range of [ x-deltat, x+deltat ], as an abnormal point pair, and deltat is a preset error threshold value;
the missing saw blade teeth are identified based on the outlier pairs.
2. The method for determining error in sawing process of aluminum profile based on data analysis according to claim 1, wherein the filtering the membership sub-gaussian model of the pixel point at the same position in each frame of saw blade gray level image according to the membership degree comprises the following specific steps:
and for the pixel point of each position in each frame of saw blade gray level image, acquiring a sub-Gaussian model corresponding to the largest membership degree in the membership degrees of the pixel point to each sub-Gaussian model of the position, and taking the sub-Gaussian model as the membership sub-Gaussian model of the pixel point.
3. The method for determining error in sawing process of aluminum profile based on data analysis according to claim 1, wherein the step of obtaining the membership sequence of each position in the saw blade gray level image of the continuous frame according to the membership sub-gaussian model of the pixel point of the same position in the saw blade gray level image of each frame comprises the following specific steps:
and forming a one-dimensional sequence according to the sequence of the membership sub Gaussian model of the pixel point at the same position in all frame saw blade gray images and taking the sequence of the sequence as the membership sequence of the position.
4. The method for determining errors in the sawing process of the aluminum profile based on data analysis according to claim 1, wherein the method for obtaining the difference function is as follows:
wherein l is the difference between the frame numbers of the ith frame of saw blade gray level image and the (i+l) th frame of saw blade gray level image; exp () represents an exponential function based on a natural constant; the absolute value symbol is denoted by i.
5. The error judging method for the sawing process of the aluminum profile based on data analysis according to claim 1, wherein the identifying of the missing saw blade teeth according to the abnormal point pairs comprises the following specific steps:
and presetting a quantity threshold, and when the quantity of the abnormal point pairs is larger than the quantity threshold, missing saw blade teeth.
6. The error judgment method for the sawing process of the aluminum profile based on data analysis according to claim 1, wherein the step of screening abnormal pixels according to the abnormality degree comprises the following specific steps:
presetting an abnormal threshold value, and taking a pixel point with the degree of abnormality greater than the abnormal threshold value in each frame of saw blade gray level image as an abnormal pixel point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311441203.XA CN117173208B (en) | 2023-11-01 | 2023-11-01 | Error judgment method for sawing process of aluminum profile based on data analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311441203.XA CN117173208B (en) | 2023-11-01 | 2023-11-01 | Error judgment method for sawing process of aluminum profile based on data analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117173208A CN117173208A (en) | 2023-12-05 |
| CN117173208B true CN117173208B (en) | 2024-03-12 |
Family
ID=88943486
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311441203.XA Active CN117173208B (en) | 2023-11-01 | 2023-11-01 | Error judgment method for sawing process of aluminum profile based on data analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117173208B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103761727A (en) * | 2013-12-28 | 2014-04-30 | 辽宁师范大学 | Robust image segmentation method based on self-adaption Gaussian mixture model |
| CN112991362A (en) * | 2021-03-17 | 2021-06-18 | 合肥高晶光电科技有限公司 | Color sorter adhesion material image segmentation method based on Gaussian mixture model |
| CN114187286A (en) * | 2021-12-17 | 2022-03-15 | 沭阳县旭东木业制品厂 | Wood plate surface machining quality control method based on machine vision |
| CN114897898A (en) * | 2022-07-13 | 2022-08-12 | 江苏绿泉装饰工程有限公司 | Wood board quality classification method based on image processing |
| CN114972203A (en) * | 2022-04-29 | 2022-08-30 | 南通市立新机械制造有限公司 | Mechanical part rolling abnormity detection method based on watershed segmentation |
| CN115601364A (en) * | 2022-12-14 | 2023-01-13 | 惠州威尔高电子有限公司（Cn） | Golden finger circuit board detection method based on image analysis |
| CN116542972A (en) * | 2023-07-04 | 2023-08-04 | 山东阁林板建材科技有限公司 | Wall plate surface defect rapid detection method based on artificial intelligence |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9217694B2 (en) * | 2003-10-21 | 2015-12-22 | Leica Microsystems Cms Gmbh | Method for automatically generating laser cutting lines in laser microdissection processes |
| CN116823826B (en) * | 2023-08-29 | 2023-11-03 | 无锡康贝电子设备有限公司 | Numerical control machine tool tipping abnormity detection method |
-
2023
- 2023-11-01 CN CN202311441203.XA patent/CN117173208B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103761727A (en) * | 2013-12-28 | 2014-04-30 | 辽宁师范大学 | Robust image segmentation method based on self-adaption Gaussian mixture model |
| CN112991362A (en) * | 2021-03-17 | 2021-06-18 | 合肥高晶光电科技有限公司 | Color sorter adhesion material image segmentation method based on Gaussian mixture model |
| CN114187286A (en) * | 2021-12-17 | 2022-03-15 | 沭阳县旭东木业制品厂 | Wood plate surface machining quality control method based on machine vision |
| CN114972203A (en) * | 2022-04-29 | 2022-08-30 | 南通市立新机械制造有限公司 | Mechanical part rolling abnormity detection method based on watershed segmentation |
| CN114897898A (en) * | 2022-07-13 | 2022-08-12 | 江苏绿泉装饰工程有限公司 | Wood board quality classification method based on image processing |
| CN115601364A (en) * | 2022-12-14 | 2023-01-13 | 惠州威尔高电子有限公司（Cn） | Golden finger circuit board detection method based on image analysis |
| CN116542972A (en) * | 2023-07-04 | 2023-08-04 | 山东阁林板建材科技有限公司 | Wall plate surface defect rapid detection method based on artificial intelligence |
Non-Patent Citations (2)
| Title |
|---|
| 3D shape measurement based on the unequal-period combination of shifting Gray code and dual-frequency phase-shifting fringes;Shuang Yu et al;《Optics Communications》;1-12 * |
| 基于视觉的皮革自适应切割技术研究_卢孔宝2019年第02期;卢孔宝;《中国优秀硕士学位论文全文数据库（电子期刊）》;第2019卷(第02期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117173208A (en) | 2023-12-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110084165B (en) | Intelligent identification and early warning method for abnormal events in open scene of power field based on edge calculation | |
| US20230306577A1 (en) | Cross-scale defect detection method based on deep learning | |
| CN112906463A (en) | Image-based fire detection method, device, equipment and storage medium | |
| CN110765951A (en) | Remote sensing image airplane target detection method based on bounding box correction algorithm | |
| CN113469951B (en) | Hub defect detection method based on cascade region convolutional neural network | |
| CN107248174A (en) | A kind of method for tracking target based on TLD algorithms | |
| CN115065708B (en) | Industrial Internet of things system based on machine vision detection and control method thereof | |
| CN113421192B (en) | Training method of object statistical model, and statistical method and device of target object | |
| EP3605453A1 (en) | Convolutional neural network based inspection of blade-defects of a wind turbine | |
| CN115019181B (en) | Remote sensing image rotating target detection method, electronic equipment and storage medium | |
| CN110728269B (en) | High-speed rail contact net support pole number plate identification method based on C2 detection data | |
| CN117173208B (en) | Error judgment method for sawing process of aluminum profile based on data analysis | |
| Kähler et al. | Anomaly detection for industrial surface inspection: application in maintenance of aircraft components | |
| CN110837760B (en) | Target detection method, training method and device for target detection | |
| CN114707577A (en) | Anomaly detection method and system based on self-confrontation variational self-encoder | |
| CN112801959B (en) | Auxiliary assembly system based on visual feature recognition | |
| CN111612681A (en) | Data acquisition method, watermark identification method, watermark removal method and device | |
| CN113077423A (en) | Laser selective melting pool image analysis system based on convolutional neural network | |
| CN113034359A (en) | Intelligent diagnosis method and device for generator set stator and rotor defects, computer equipment and storage medium | |
| CN113901944B (en) | Marine organism target detection method based on improved YOLO algorithm | |
| CN106447045A (en) | Evaluation method for ADAS (Advanced Driver Assistant System) based on machine learning | |
| CN115797314A (en) | Part surface defect detection method, system, equipment and storage medium | |
| CN111753588A (en) | Power equipment online perception and focusing algorithm based on deep learning | |
| CN111476129A (en) | Soil impurity detection method based on deep learning | |
| CN108181315B (en) | Image processing-based biscuit damage detection device and detection method |
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 |