CN115265669B - A quality inspection system for pipe cutting hot melt process based on binary classifier - Google Patents
A quality inspection system for pipe cutting hot melt process based on binary classifier Download PDFInfo
- Publication number
- CN115265669B CN115265669B CN202211133779.5A CN202211133779A CN115265669B CN 115265669 B CN115265669 B CN 115265669B CN 202211133779 A CN202211133779 A CN 202211133779A CN 115265669 B CN115265669 B CN 115265669B
- Authority
- CN
- China
- Prior art keywords
- cutting
- hot melting
- welding cooling
- working
- pipe
- 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
- 238000005520 cutting process Methods 0.000 title claims abstract description 112
- 238000000034 method Methods 0.000 title claims abstract description 30
- 239000012943 hotmelt Substances 0.000 title abstract description 8
- 230000008569 process Effects 0.000 title abstract description 8
- 238000007689 inspection Methods 0.000 title 1
- 238000001816 cooling Methods 0.000 claims abstract description 118
- 238000003466 welding Methods 0.000 claims abstract description 91
- 238000002844 melting Methods 0.000 claims abstract description 75
- 230000002159 abnormal effect Effects 0.000 claims abstract description 68
- 230000008018 melting Effects 0.000 claims abstract description 66
- 239000013598 vector Substances 0.000 claims abstract description 59
- 238000001514 detection method Methods 0.000 claims abstract description 26
- 230000004044 response Effects 0.000 claims description 36
- 238000010309 melting process Methods 0.000 claims description 22
- 230000004927 fusion Effects 0.000 claims description 16
- 230000008859 change Effects 0.000 claims description 9
- 238000005259 measurement Methods 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 4
- 230000000694 effects Effects 0.000 description 6
- 210000001503 joint Anatomy 0.000 description 4
- 238000003801 milling Methods 0.000 description 4
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 4
- 238000012545 processing Methods 0.000 description 3
- 238000012790 confirmation Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000010438 heat treatment Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 235000014435 Mentha Nutrition 0.000 description 1
- 241001072983 Mentha Species 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 235000014569 mints Nutrition 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- 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
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Investigating Or Analyzing Materials Using Thermal Means (AREA)
Abstract
本发明涉及塑料管材连接技术领域，具体涉及一种基于二分类器的管材切削热熔工艺质量检测系统，该系统包括第一数据采集单元：用于采集切削热熔工作中管道管口的直径、冷却温度和端面图像，得到管道管口的形变量典型程度、熔接冷却稳定性和端面外轮廓对应的傅里叶描述子；第二数据采集单元：获取切削热熔工作的工作特例程度和管口端面切削特征，以结合形变量典型程度和熔接冷却稳定性组成工作质量特征向量；异常状态检测单元：利用多个工作质量特征向量训练Adaboost二分类器，由训练好的Adaboost二分类器对切削热熔工作进行异常状态检测。本发明能够快速准确地进行异常预警，以方便及时发现问题，对工艺质量检测起到辅助作用。
The invention relates to the technical field of plastic pipe connection, and specifically relates to a pipe cutting hot melt process quality detection system based on a two-classifier. The system includes a first data acquisition unit: used to collect the diameter of the pipe nozzle during cutting hot melt work, The cooling temperature and end face image are used to obtain the typical degree of deformation of the pipe nozzle, the welding cooling stability and the Fourier descriptor corresponding to the outer contour of the end face; the second data acquisition unit: obtain the work special degree and nozzle of the cutting hot melt work The end face cutting characteristics are combined with the typical degree of deformation and welding cooling stability to form a work quality feature vector; the abnormal state detection unit: uses multiple work quality feature vectors to train an Adaboost binary classifier, and the trained Adaboost binary classifier is used to detect the cutting heat Detect abnormal conditions during melting work. The invention can quickly and accurately perform abnormal early warning to facilitate timely discovery of problems and play an auxiliary role in process quality detection.
Description
Technical Field
The invention relates to the technical field of plastic pipe connection, in particular to a pipe cutting and hot melting process quality detection system based on two classifiers.
Background
At present, the pipe is cut firstly in the cutting hot melting process of the pipe, and then the cut pipe is connected in a hot melting mode. And the quality of the cutting and hot melting process largely determines the mounting quality of the pipe. The quality evaluation of the cutting and hot melting process of the pipe mainly depends on appearance detection, but when the cutting and hot melting process is abnormal, abnormal conditions such as false welding and the like can occur, and the abnormal conditions can not be found through the appearance detection, so that a great error can occur in the evaluation result of the quality of the cutting and hot melting process of the pipe.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a pipe cutting hot melting process quality detection system based on two classifiers, which adopts the following technical scheme:
the first data acquisition unit is used for acquiring the diameter of the pipe orifice of the pipe based on different measurement angles in the current cutting and hot melting operation to obtain a diameter sequence, and obtaining the typical degree of deformation of the pipe orifice of the pipe according to the diameter difference in the diameter sequence; based on the sampling frequency, acquiring a temperature change sequence when the cooling temperature of each region of the pipe orifice reaches a temperature threshold value, respectively calculating a welding cooling coefficient corresponding to each temperature change sequence to obtain a welding cooling coefficient sequence, and obtaining the welding cooling stability of the pipe orifice according to the difference of the welding cooling coefficients in the welding cooling coefficient sequence; acquiring an end face image of a pipe orifice of the pipeline, and acquiring a Fourier descriptor corresponding to the outer contour of the end face according to the end face image;
the second data acquisition unit is used for combining the welding cooling coefficient sequence, the welding cooling stability and the Fourier descriptor to calculate the working special degree of the current cutting hot melting work; forming a low-frequency component in the Fourier descriptor into a pipe orifice end face cutting characteristic under the current cutting and hot melting work; forming the deformation typical degree, the welding cooling stability, the working special case degree and the pipe orifice end face cutting characteristic into a working quality characteristic vector of the current cutting and hot melting work;
the abnormal state detection unit is used for acquiring the working quality feature vectors of a plurality of cutting hot melting works, and clustering all the working quality feature vectors according to the difference between the working special degrees to obtain a normal sample set and an abnormal sample set which are formed by the working quality feature vectors; training the Adaboost two-classifier by using a normal sample set and an abnormal sample set to obtain a strong classifier of the Adaboost two-classifier, and detecting abnormal states of the cutting hot melting work by using the strong classifier.
Further, the method for obtaining the deformation typical degree of the pipe orifice of the pipeline according to the diameter difference in the diameter sequence in the first data acquisition unit comprises the following steps:
and respectively calculating the absolute value of the difference between any two diameters in the diameter sequence, calculating the variance of the absolute value of the difference according to the absolute value of the difference, and taking the variance as the typical degree of the deformation.
Further, the method for obtaining the welding cooling stability of the pipe orifice according to the difference of the welding cooling coefficients in the welding cooling coefficient sequence in the first data acquisition unit includes:
respectively calculating any two of the welding cooling coefficient sequencesCalculating the weld cooling stability based on the ratio between the weld cooling coefficients, wherein the weld cooling stabilityThe calculation formula of (2) is as follows:
wherein,,is a mean function; />As a function of absolute value; />For the +.>A plurality of fusion cooling coefficients; />For the +.>A plurality of fusion cooling coefficients; />Is constant.
Further, the method for obtaining the degree of the working special case in the second data acquisition unit comprises the following steps:
obtaining a standard Fourier descriptor corresponding to the outer contour of the standard end face of the pipe orifice of the pipeline, and calculating the similarity between the Fourier descriptor and the standard Fourier descriptor; obtaining an average welding cooling coefficient and a maximum welding cooling coefficient of the welding cooling coefficient sequence, and obtaining a working special case degree by combining the similarity degree, the average welding cooling coefficient, the maximum welding cooling coefficient and the welding cooling stability, wherein a calculation formula of the working special case degree is as follows:
wherein,,the degree of the working special case is the degree of the working special case; cosine is a Cosine similarity function; mean is the mean function; max is a maximum function; />Is a standard fourier descriptor; />Is the fourier descriptor; />Cooling stability for the weld; />A sequence of cooling coefficients for said fusion.
Further, the method for obtaining the normal sample set and the abnormal sample set in the abnormal state detection unit includes:
respectively calculating the difference value corresponding to the working special case degree in any two working quality feature vectors, taking the difference value as a sample distance, and performing DBSCAN clustering on all the working quality feature vectors based on the sample distance to obtain a plurality of clusters and isolated points;
and forming the working quality feature vectors corresponding to the isolated points into the abnormal sample set, and forming the working quality feature vectors in all the cluster clusters into the normal sample set.
Further, the abnormal state detection unit further includes an optimization for the abnormal sample set, and the optimization method includes:
and counting the working special degree in the abnormal sample set to obtain a median value of the working special degree, obtaining a target working quality characteristic vector corresponding to the working special degree, which is larger than the median value, in the abnormal sample set, randomly combining all elements in the target working quality characteristic vector to obtain a plurality of new working quality characteristic vectors, and putting the new working quality characteristic vectors into the abnormal sample set.
Further, the method for detecting the abnormal state of the cutting hot melting work by using the strong classifier in the abnormal state detection unit comprises the following steps:
respectively acquiring historical response values of a plurality of historical cutting hot melting works according to the strong classifier, and acquiring a response value threshold according to the historical response values;
and respectively acquiring a real-time response value of the real-time cutting and hot melting work and a response value of the continuous K times of cutting and hot melting work before the real-time cutting and hot melting work according to the strong classifier, wherein K is a positive integer, and performing abnormal early warning of the real-time cutting and hot melting work when the real-time response value and the response value are lower than the response value threshold.
The embodiment of the invention has at least the following beneficial effects: acquiring the diameter of a pipe orifice of a pipeline in cutting hot melting work, welding cooling coefficients of all areas and an end face outline to obtain a working quality feature vector of the cutting hot melting work, dividing the working quality feature vector of a plurality of cutting hot melting works into a normal sample and an abnormal sample, and training an Adaboost two classifier to obtain the Adaboost two classifier for intelligently monitoring the working state of the cutting hot melting work; the working quality feature vector of the real-time cutting and hot melting work is used for obtaining the response value of the real-time cutting and hot melting work through the trained Adaboost two classifier, and the abnormal early warning can be rapidly and accurately carried out according to the response value, so that problems can be conveniently and timely found out, the auxiliary effect of the cutting and hot melting work is achieved, and the quality of the cutting and hot melting process is improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a block diagram of a pipe cutting hot melting process quality detection system based on two classifiers according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following is a detailed description of specific implementation, structure, characteristics and effects of the pipe cutting hot melting process quality detection system based on two classifiers according to the invention with reference to 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 invention provides a concrete scheme of a pipe cutting hot melting process quality detection system based on two classifiers, which is specifically described below with reference to the accompanying drawings.
Referring to fig. 1, a block diagram of a pipe cutting hot melting process quality detection system based on two classifiers according to an embodiment of the present invention is shown, where the system includes: a first data acquisition unit 10, a second data acquisition unit 20, and an abnormal state detection unit 30.
The first data acquisition unit 10 is used for acquiring the diameter of the pipe orifice of the pipe based on different measurement angles in the current cutting and hot melting operation to obtain a diameter sequence, and obtaining the typical deformation degree of the pipe orifice of the pipe according to the diameter difference in the diameter sequence; based on the sampling frequency, acquiring a temperature change sequence when the cooling temperature of each region of the pipe orifice reaches a temperature threshold value, respectively calculating a welding cooling coefficient corresponding to each temperature change sequence to obtain a welding cooling coefficient sequence, and obtaining the welding cooling stability of the pipe orifice according to the difference of the welding cooling coefficients in the welding cooling coefficient sequence; and acquiring an end face image of the pipe orifice of the pipe, and acquiring a Fourier descriptor corresponding to the outer contour of the end face according to the end face image.
Specifically, because the pipeline is when being clamped by the clamp, the deformation of the pipe orifice can be caused due to overlarge clamping force, thereby affecting the butt joint precision of the pipeline, the deformation information of the pipe orifice of the pipeline in the cutting hot melting work is acquired, and the acquisition method of the deformation information is as follows: because the pipeline is when being held by anchor clamps, the orificial diameter of pipeline can embody its deformation, consequently this scheme gathers direct measurement method, utilizes line scanning laser rangefinder to carry out direct measurement to the orificial diameter of pipeline, and carries out diameter measurement along the orificial different angles of pipeline to constitute diameter sequence D with the diameter under the different measurement angles.
It should be noted that, because the line scanning laser ranging device is widely applied in industrial machine vision, the process of how the line scanning laser ranging device obtains the diameter of the pipe orifice in the scheme is not repeated.
And respectively calculating the absolute value of the difference between any two diameters in the diameter sequence D, calculating the variance of the absolute value of the difference according to the absolute value of the difference, and taking the variance as the deformation typical degree P of the pipe orifice of the pipeline.
Because water possibly remains at the lower part of the water pipe, the water absorbs heat during hot melting, the welding cooling time of the lower part of the pipe is faster than that of other parts, and then the false welding condition of false welding caused by insufficient welding time can occur, and water leakage can be caused during later use; and the cooling time of the pipeline hot melt surface is too fast in winter, the phenomenon that the joint of the two pipe surfaces is not completely welded and is cooled can occur, and the condition of false welding and false welding of the pipeline can also occur, so that the cooling time of the pipeline is analyzed to determine the welding condition in the cutting hot melt work.
The scheme is used for collecting cooling time, the point type infrared thermometer is used for collecting cooling time, and after heating of the pipe orifice of the pipeline is completed from the heating plate, the temperature of the pipe orifice of the pipeline is detected. And setting a cooling temperature threshold, and considering that the welding is completed when the cooling temperature reaches the temperature threshold.
And taking 0.1S as sampling frequency, acquiring a temperature composition temperature change sequence M at each sampling moment based on the sampling frequency in the process that the cooling temperature of the current area of the pipe orifice of the pipe reaches a temperature threshold value, and acquiring the temperature change sequence of each area of the pipe orifice of the pipe in the same way.
And calculating the temperature difference value of each temperature change sequence by using a range function, and taking the temperature difference value as a welding cooling coefficient of a corresponding region of the pipe orifice of the pipeline to form a welding cooling coefficient sequence V.
Under normal conditions, the welding cooling coefficient of the welding cooling device can be stabilized at a value, if the welding cooling coefficient is smaller than the steady state, the abnormal operation of the welding cooling device can be described to a certain extent, and the abnormal appearance of the form is possible at the moment, so that the quality of the hot-melting butt joint of the whole pipeline is affected, and the welding cooling stability of the pipe orifice of the pipeline is obtained according to the difference of the welding cooling coefficients in the welding cooling coefficient sequence, and the obtaining method of the welding cooling stability is as follows: calculating the ratio between any two welding cooling coefficients in the welding cooling coefficient sequence, and calculating the welding cooling stability of the pipe orifice of the pipeline based on the ratio, wherein the welding cooling stabilityThe calculation formula of (2) is as follows:
wherein,,for averaging functions, for obtaining processed dataAverage size; />The absolute value function is used for taking the processed data to a non-negative interval; />For welding cooling coefficient sequences +.>A plurality of fusion cooling coefficients;for welding cooling coefficient sequences +.>A plurality of fusion cooling coefficients; />Is constant.
When the two welding cooling coefficients are the same, the index obtained by processing is 1, if the two welding cooling coefficients are not equal, the index is smaller than 1, and the stability of welding cooling can be directly judged through the size of an index function; when the average size after processing in the whole record is 1, that isAt a value of 1, it is indicated that the weld cooling coefficient sequence does not fluctuate, i.e., it is stable.
Since milling is an important step of hot melting, the end face is required to be flat, smooth and vertical after milling, otherwise, the butt joint precision of the pipe orifice during hot melting is affected. However, when the temperature in winter is extremely low and the pipe reaches a brittle point, the milling cutter is used for milling, so that the pipe is easy to crack, the hot melting effect is affected at the moment, when the temperature in summer is extremely hot, the pipe is possibly softened, when the clamp clamps the pipe, the clamping part possibly causes insufficient roundness of the pipe orifice due to the generated deformation, and the butt joint precision of the pipe orifice hot melting is also affected, so that the cutting surface information of the pipe orifice of the pipeline in the hot melting cutting work is required to be obtained.
The method comprises the steps of collecting cutting surface information by using an industrial camera shooting mode, namely shooting an end surface of a pipe orifice of a pipe to obtain an end surface image of the pipe orifice of the pipe, performing binary segmentation on the end surface image by using a conventional machine vision technology to obtain an end surface outline of the pipe orifice of the pipe, and analyzing frequency components of the outline by using a Fourier descriptor based on the end surface outline to obtain a Fourier descriptor F of the end surface outline.
The diameter sequence D, the deformation typical degree P, the welding cooling coefficient sequence V and the welding cooling stability of a pipe orifice in a cutting hot-melting work are adoptedAnd a Fourier descriptor F of the outer contour of the end face form working characteristic data of the cutting hot melting work.
A second data acquisition unit 20 for calculating the degree of working special cases of the current cutting hot-melt work by combining the welding cooling coefficient sequence, the welding cooling stability and the fourier descriptor; the low-frequency component in the Fourier descriptor is formed into a pipe orifice end face cutting characteristic under the current cutting and hot melting work; and forming the deformation typical degree, welding cooling stability, working special case degree and pipe orifice end face cutting characteristics into a working quality characteristic vector of the current cutting and hot melting work.
Specifically, when the cutting hot-melting work is abnormal, the external contour of the end face of the pipe orifice of the pipe and the welding cooling coefficient are obviously changed, so that the working special case degree U of the cutting hot-melting work is determined according to the working characteristic data of the cutting hot-melting work, and the method for acquiring the working special case degree U is as follows: obtaining a standard Fourier descriptor corresponding to the outer contour of the standard end face of the pipe orifice of the pipeline by usingThe similarity degree between the Fourier descriptor and the standard Fourier descriptor is calculated through the function; simultaneously obtaining the average fusion cooling coefficient mean (V) and the average fusion cooling coefficient mean (V) of the fusion cooling coefficient sequence VMaximum fusion cooling coefficient max (V), combined with similarity, average fusion cooling coefficient mean (V), maximum fusion cooling coefficient max (V), and fusion cooling stability->And obtaining the working special case degree U corresponding to the cutting and hot melting work.
As an example, the calculation formula of the work special case degree U is:
wherein Cosine is a Cosine similarity function for comparing the degree of similarity between two vectors; mean is the mean function; max is a maximum function;the standard Fourier descriptor is corresponding to the standard end surface outline of the pipe orifice.
It should be noted that, when the fourier descriptor of the cutting and hot melting work is more similar to the standard fourier descriptor, the degree of similarity is larger, and the corresponding degree of working special cases is smaller; when the stability of the cooling temperature of the cutting hot melting work is larger, namely the welding cooling stability is larger, the corresponding working special degree is smaller; when there is a significant difference between the average welding cooling coefficient mean (V) and the maximum welding cooling coefficient max (V) of the welding cooling coefficient series V, then the ratioSmaller and greater the corresponding work specials.
Since the fourier descriptor in the working characteristic data of the cutting hot-melt work contains a relatively large amount of information and part of the information is formed by errors, the high-frequency component of the fourier descriptor is useless, and the fourier descriptor F is processed: extracting low-frequency components in the Fourier descriptor, forming the low-frequency components into a pipe orifice end face cutting feature W under the cutting and hot melting work, and using the low-frequency components to represent the low-frequency components in fewer vectors so as to reduce subsequent calculation amount.
Preferably, the first 5 low frequency components are selected to form the nozzle end face cutting feature W in the embodiment of the present invention due to the variable number of fourier descriptors.
The purpose of using the dimension reduction vector of the fourier descriptor to represent the end face cutting characteristics of the pipe orifice is to reduce the calculation amount when the Adaboost two-classifier is built, improve the judgment performance and avoid the error of end face image acquisition.
The deformation typical degree P of the pipe orifice in the cutting hot melting work and the welding cooling stability are adoptedDegree of work special case->And the pipe orifice end face cutting characteristic W forms a working quality characteristic vector X= { Q, P, U, W } corresponding to the cutting hot melting work.
An abnormal state detection unit 30, configured to obtain a plurality of working quality feature vectors of the cutting and hot melting operation, and cluster all the working quality feature vectors according to differences between the degrees of the working special cases, so as to obtain a normal sample set and an abnormal sample set that are composed of the working quality feature vectors; training the Adaboost two-classifier by using the normal sample set and the abnormal sample set to obtain a strong classifier of the Adaboost two-classifier, and detecting abnormal states of the cutting hot melting work by using the strong classifier.
Specifically, since the abnormal condition of a single cutting and hot-melting work is relatively single and the probability of occurrence of the abnormality is relatively small, the working quality feature vectors X of a plurality of cutting and hot-melting works are obtained by using the methods of step S001 and step S002.
Cutting hot-melting work with similar working states is divided into a group according to the working quality characteristic vector X of each cutting hot-melting work so as to determine a plurality of working states. The dividing method comprises the following steps: respectively calculating any two working qualitiesCorresponding working special case degree in feature vector XTaking the difference value as a sample distance, and performing DBSCAN clustering on all the working quality feature vectors X based on the sample distance, wherein the searching radius eps in the DBSCAN clustering is defaulted to be 0.5, and the minimum value mints in the clusters is set to be 4, so as to obtain a plurality of clustering clusters and isolated points.
In the embodiment of the invention, the working quality feature vectors X corresponding to the isolated points are used as abnormal samples, so that the working quality feature vectors X corresponding to all the isolated points form an abnormal sample set, each working quality feature vector X in the cluster is used as a normal sample, and the working quality feature vectors X in all the cluster form a normal sample set.
And taking the working quality feature vector X as input data of the Adaboost two-classifier, and training the Adaboost two-classifier by utilizing the abnormal sample set and the normal sample set to determine a strong classifier of the Adaboost two-classifier.
Because the number of the abnormal samples is small, the states of all weak classifiers in the Adaboost two classifiers may not be effectively constrained, so that the abnormal sample set is expanded according to the working quality feature vector X in the abnormal sample set: because the abnormal occurrence of the cutting hot melting work can be obviously determined under the condition that the work special degree U is larger, the work special degree U in the abnormal sample set is counted to obtain the median of the work special degreeThen, for the working special cases in the abnormal sample set, the degree is larger than the median +.>The target working quality characteristic vector of the (2) is expanded, namely the deformation typical degree P and welding cooling stability in the target working quality characteristic vector are expanded>Degree of work special case->And the pipe orifice end face cutting feature W is randomly combined to obtain a plurality of new working quality feature vectors, and the new working quality feature vectors are integrated into an abnormal sample set to optimize the abnormal sample set.
According to the strong classifier, historical response values of continuous M historical cutting hot melting works are respectively obtained, M is a positive integer, and a response value threshold is obtained according to the historical response values: the response value of the strong classifier is larger than 0, which indicates that the corresponding working quality feature vector belongs to the feature vector in the normal working state, otherwise, the response value is smaller than 0, which indicates that the corresponding working quality feature vector belongs to the feature vector in the abnormal working state; and calculating the average value of the response values of Top-20% in the M historical response values, and taking the average value of the response values as a response value threshold.
The method comprises the steps of respectively obtaining real-time response values of working quality feature vectors of real-time cutting and hot melting work and response values of the working quality feature vectors of K cutting and hot melting works before the real-time cutting and hot melting work by using a strong classifier, immediately carrying out abnormal early warning of the real-time cutting and hot melting work when the real-time response values and the response values are lower than a response value threshold, and adopting corresponding measures to carry out abnormal confirmation, wherein the method comprises the following specific steps: when the response values of the K cutting hot melting operations are not all smaller than the response value threshold value, an operator is enabled to check the operation effect, but the fact that the problem does exist is not meant; when the response values of the K real-time cutting and hot melting operations are smaller than the threshold value of the response values, the occurrence of problems is confirmed, and the pipeline is directly treated as waste.
It should be noted that, based on the logic of abnormal confirmation, an operator can combine the site situation, determine whether the cutting hot melting is abnormal or nearly abnormal based on the output result of the Adaboost two-classifier, so that energy is saved in the repeated cutting hot melting process operation process, and the Adaboost two-classifier prompts which situations need to be noted.
The reason for monitoring the working state of the cutting hot melting work by using the Adaboost two classifier is as follows: the two classification effects of Adaboost are rough classification, and the conditions of cutting hot melting are changeable and the working characteristics, namely the clamps and cutters are easy to accumulate heat, so that the end face, the cooling coefficient, the deformation and the like of the pipe orifice of the pipeline are changed slowly, and the combined judgment is carried out by combining the working states of a plurality of cutting hot melting works, so that the abnormal judgment result is more accurate.
In summary, the embodiment of the invention provides a pipe cutting hot melting process quality detection system based on two classifiers, which comprises a first data acquisition unit: the method is used for collecting the diameter of a pipe orifice of a pipeline in a cutting hot melting work, the cooling temperature of each area and an end face image, and analyzing to obtain Fourier descriptors corresponding to the deformation typical degree, the welding cooling stability and the end face outline of the pipe orifice; a second data acquisition unit: the method comprises the steps of obtaining the working special degree of cutting hot melting work and the cutting characteristic of the end face of a pipe orifice, and combining the deformation typical degree and the welding cooling stability to form a working quality characteristic vector; abnormal state detection means: and training Adaboost two classifiers by using a plurality of working quality feature vectors, and detecting abnormal states of the cutting hot melting work by the trained Adaboost two classifiers. The invention can rapidly and accurately perform abnormal early warning so as to conveniently and timely find problems and play an auxiliary role in process quality.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. And the foregoing description has been directed to specific embodiments of this specification. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
The foregoing description of the preferred embodiments of the present invention is not intended to be limiting, but rather, any modifications, equivalents, improvements, etc. that fall within the principles of the present invention are intended to be included within the scope of the present invention.
Claims (6)
1. A pipe cutting hot melting process quality detection system based on two classifiers is characterized in that the system comprises:
the first data acquisition unit is used for acquiring the diameter of the pipe orifice of the pipe based on different measurement angles in the current cutting and hot melting operation to obtain a diameter sequence, and obtaining the typical degree of deformation of the pipe orifice of the pipe according to the diameter difference in the diameter sequence; based on the sampling frequency, acquiring a temperature change sequence when the cooling temperature of each region of the pipe orifice reaches a temperature threshold value, respectively calculating a welding cooling coefficient corresponding to each temperature change sequence to obtain a welding cooling coefficient sequence, and obtaining the welding cooling stability of the pipe orifice according to the difference of the welding cooling coefficients in the welding cooling coefficient sequence; acquiring an end face image of a pipe orifice of the pipeline, and acquiring a Fourier descriptor corresponding to the outer contour of the end face according to the end face image;
the second data acquisition unit is used for combining the welding cooling coefficient sequence, the welding cooling stability and the Fourier descriptor to calculate the working special degree of the current cutting hot melting work; forming a low-frequency component in the Fourier descriptor into a pipe orifice end face cutting characteristic under the current cutting and hot melting work; forming the deformation typical degree, the welding cooling stability, the working special case degree and the pipe orifice end face cutting characteristic into a working quality characteristic vector of the current cutting and hot melting work;
the abnormal state detection unit is used for acquiring the working quality feature vectors of a plurality of cutting hot melting works, and clustering all the working quality feature vectors according to the difference between the working special degrees to obtain a normal sample set and an abnormal sample set which are formed by the working quality feature vectors; training the Adaboost two-classifier by using a normal sample set and an abnormal sample set to obtain a strong classifier of the Adaboost two-classifier, and detecting abnormal states of cutting hot melting work by using the strong classifier;
the method for acquiring the working special case degree in the second data acquisition unit comprises the following steps:
obtaining a standard Fourier descriptor corresponding to the outer contour of the standard end face of the pipe orifice of the pipeline, and calculating the similarity between the Fourier descriptor and the standard Fourier descriptor; obtaining an average welding cooling coefficient and a maximum welding cooling coefficient of the welding cooling coefficient sequence, and obtaining a working special case degree by combining the similarity degree, the average welding cooling coefficient, the maximum welding cooling coefficient and the welding cooling stability, wherein a calculation formula of the working special case degree is as follows:
wherein U is the working special case degree; cosine is a Cosine similarity function; mean is the mean function; max is a maximum function; f (F) normal Is a standard fourier descriptor; f is the Fourier descriptor; q is the welding cooling stability; v is the welding cooling coefficient sequence.
2. The system for detecting the quality of a pipe cutting and hot melting process based on two classifiers according to claim 1, wherein the method for obtaining the deformation typical degree of the pipe orifice of the pipe according to the diameter difference in the diameter sequence in the first data acquisition unit comprises the following steps:
and respectively calculating the absolute value of the difference between any two diameters in the diameter sequence, calculating the variance of the absolute value of the difference according to the absolute value of the difference, and taking the variance as the typical degree of the deformation.
3. The system for detecting the quality of a pipe cutting hot melting process based on two classifiers according to claim 1, wherein the method for obtaining the welding cooling stability of the pipe orifice according to the difference of the welding cooling coefficients in the welding cooling coefficient sequence in the first data acquisition unit comprises the following steps:
and respectively calculating the ratio between any two welding cooling coefficients in the welding cooling coefficient sequence, and calculating the welding cooling stability based on the ratio, wherein the calculation formula of the welding cooling stability Q is as follows:
wherein mean is a mean function; abs is an absolute function; v (V) i An i-th fusion cooling coefficient in the sequence of fusion cooling coefficients; v (V) j A j-th fusion cooling coefficient in the fusion cooling coefficient sequence; e is a constant.
4. The system for detecting the quality of a pipe cutting hot melting process based on two classifiers according to claim 1, wherein the method for acquiring a normal sample set and an abnormal sample set in the abnormal state detection unit comprises the following steps:
respectively calculating the difference value corresponding to the working special case degree in any two working quality feature vectors, taking the difference value as a sample distance, and performing DBSCAN clustering on all the working quality feature vectors based on the sample distance to obtain a plurality of clusters and isolated points;
and forming the working quality feature vectors corresponding to the isolated points into the abnormal sample set, and forming the working quality feature vectors in all the cluster clusters into the normal sample set.
5. The pipe cutting hot melting process quality detection system based on two classifiers according to claim 1, wherein the abnormal state detection unit further comprises optimization of the abnormal sample set, and the optimization method comprises the following steps:
and counting the working special degree in the abnormal sample set to obtain a median value of the working special degree, obtaining a target working quality characteristic vector corresponding to the working special degree, which is larger than the median value, in the abnormal sample set, randomly combining all elements in the target working quality characteristic vector to obtain a plurality of new working quality characteristic vectors, and putting the new working quality characteristic vectors into the abnormal sample set.
6. The system for detecting the quality of a pipe cutting and hot melting process based on two classifiers according to claim 1, wherein the method for detecting the abnormal state of the cutting and hot melting process by using the strong classifier in the abnormal state detection unit comprises the following steps:
respectively acquiring historical response values of a plurality of historical cutting hot melting works according to the strong classifier, and acquiring a response value threshold according to the historical response values;
and respectively acquiring a real-time response value of the real-time cutting and hot melting work and a response value of the continuous K times of cutting and hot melting work before the real-time cutting and hot melting work according to the strong classifier, wherein K is a positive integer, and performing abnormal early warning of the real-time cutting and hot melting work when the real-time response value and the response value are lower than the response value threshold.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211133779.5A CN115265669B (en) | 2022-09-19 | 2022-09-19 | A quality inspection system for pipe cutting hot melt process based on binary classifier |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211133779.5A CN115265669B (en) | 2022-09-19 | 2022-09-19 | A quality inspection system for pipe cutting hot melt process based on binary classifier |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115265669A CN115265669A (en) | 2022-11-01 |
| CN115265669B true CN115265669B (en) | 2023-10-13 |
Family
ID=83756607
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211133779.5A Active CN115265669B (en) | 2022-09-19 | 2022-09-19 | A quality inspection system for pipe cutting hot melt process based on binary classifier |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115265669B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0796433A1 (en) * | 1994-12-09 | 1997-09-24 | Exxon Chemical Patents Inc. | Plant parameter detection by monitoring of power spectral densities |
| CN107782478A (en) * | 2017-08-28 | 2018-03-09 | 南京航空航天大学 | Online pipe joint element erection stress detecting system and detection recognition method |
| CN111382773A (en) * | 2018-12-31 | 2020-07-07 | 南京拓步智能科技有限公司 | Image matching method based on nine-grid principle for monitoring inside of pipeline |
| CN114919184A (en) * | 2022-04-24 | 2022-08-19 | 南通博鎏机电科技有限公司 | PVC pipeline hot melting butt joint machine tool control method |
| CN115049600A (en) * | 2022-05-31 | 2022-09-13 | 哈尔滨工程大学 | Intelligent identification system and method for small sample pipeline defects |
-
2022
- 2022-09-19 CN CN202211133779.5A patent/CN115265669B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5790413A (en) * | 1993-03-22 | 1998-08-04 | Exxon Chemical Patents Inc. | Plant parameter detection by monitoring of power spectral densities |
| EP0796433A1 (en) * | 1994-12-09 | 1997-09-24 | Exxon Chemical Patents Inc. | Plant parameter detection by monitoring of power spectral densities |
| CN107782478A (en) * | 2017-08-28 | 2018-03-09 | 南京航空航天大学 | Online pipe joint element erection stress detecting system and detection recognition method |
| CN111382773A (en) * | 2018-12-31 | 2020-07-07 | 南京拓步智能科技有限公司 | Image matching method based on nine-grid principle for monitoring inside of pipeline |
| CN114919184A (en) * | 2022-04-24 | 2022-08-19 | 南通博鎏机电科技有限公司 | PVC pipeline hot melting butt joint machine tool control method |
| CN115049600A (en) * | 2022-05-31 | 2022-09-13 | 哈尔滨工程大学 | Intelligent identification system and method for small sample pipeline defects |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115265669A (en) | 2022-11-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110059442B (en) | Turning tool changing method based on part surface roughness and power information | |
| CN108804740B (en) | Long-distance pipeline pressure monitoring method based on integrated improved ICA-KRR algorithm | |
| CN109014544B (en) | On-line monitoring method of micro-resistance spot welding quality | |
| CN108931961A (en) | A kind of monoblock type slotting cutter worn-off damage detection method based on machine vision | |
| WO2007062563A1 (en) | On-line automatic inspection method for detecting surface flaws of steel during the pretreatment of the ship steel | |
| CN109596530B (en) | Dark field polarization imaging method for classifying pockmarks and dust defects on optical surface | |
| CN102760235B (en) | A kind of Underwater Target Classification system and method based on looking more | |
| CN111667470B (en) | Industrial pipeline flaw detection inner wall detection method based on digital image | |
| CN102279190A (en) | Image detection method for weld seam surface defects of laser welded plates of unequal thickness | |
| CN109343021B (en) | Spot mark condensation method | |
| CN118544024B (en) | Intelligent welding quality monitoring and adjusting system | |
| Chen et al. | Tool wear monitoring based on the combination of machine vision and acoustic emission | |
| CN114139772B (en) | Method for predicting workpiece surface quality based on optimized generalized regression neural network | |
| CN116664508A (en) | Weld surface quality detection method and computer readable storage medium | |
| CN101169826A (en) | On-line monitoring system for the quality of welded pipe formed by row-roller based on machine vision | |
| CN115265669B (en) | A quality inspection system for pipe cutting hot melt process based on binary classifier | |
| Atli et al. | A computer vision-based fast approach to drilling tool condition monitoring | |
| CN110738852B (en) | A detection method for intersection turn overflow based on vehicle trajectory and long short-term memory neural network | |
| CN105631857A (en) | Method and device for detecting scratch on surface of optical element | |
| CN109446964A (en) | Face detection analysis method and device based on end-to-end single-stage multiple scale detecting device | |
| CN115099266A (en) | Hard vehicle surface white layer prediction method based on gradient lifting decision tree | |
| CN115906561A (en) | On-line monitoring and quality assessment method and device for flash butt welding based on process parameters | |
| CN101780488A (en) | Internal thread cold-extrusion processing quality on-line forecasting method | |
| CN116307938B (en) | Health state assessment method for feeding system of machining center | |
| CN118966893A (en) | Smart electronic product quality safety risk monitoring and control system |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20250106 Address after: No. 8-2, Yahu Industrial Zone, Yahu Village, Tanbu Town, Huadu District, Guangzhou City, Guangdong Province 510820 Patentee after: Guangzhou nengtong Pipe Industry Co.,Ltd. Country or region after: China Address before: No. 6, Jingyi Road, Chengdong Town, Hai'an City, Nantong City, Jiangsu Province 226000 Patentee before: Bogda Intelligent Equipment (Nantong) Co.,Ltd. Country or region before: China |
|
| TR01 | Transfer of patent right |