CN118154458A - Filtering processing method of machine tool cutter image based on complex industrial environment - Google Patents
Filtering processing method of machine tool cutter image based on complex industrial environment Download PDFInfo
- Publication number
- CN118154458A CN118154458A CN202410572336.9A CN202410572336A CN118154458A CN 118154458 A CN118154458 A CN 118154458A CN 202410572336 A CN202410572336 A CN 202410572336A CN 118154458 A CN118154458 A CN 118154458A
- Authority
- CN
- China
- Prior art keywords
- analyzed
- cutter
- tool
- target
- gray
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001914 filtration Methods 0.000 title claims abstract description 80
- 238000003672 processing method Methods 0.000 title claims abstract description 18
- 238000009826 distribution Methods 0.000 claims abstract description 31
- 230000002159 abnormal effect Effects 0.000 claims abstract description 23
- 238000004891 communication Methods 0.000 claims description 122
- 238000000605 extraction Methods 0.000 claims description 48
- 238000000034 method Methods 0.000 claims description 47
- 238000010586 diagram Methods 0.000 claims description 18
- 230000003044 adaptive effect Effects 0.000 claims description 13
- 238000012216 screening Methods 0.000 claims description 12
- 230000011218 segmentation Effects 0.000 claims description 12
- 238000005520 cutting process Methods 0.000 claims description 7
- 238000001514 detection method Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 230000005856 abnormality Effects 0.000 claims description 4
- 238000003064 k means clustering Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 abstract description 8
- 239000000126 substance Substances 0.000 description 7
- 230000002452 interceptive effect Effects 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 4
- 239000002184 metal Substances 0.000 description 3
- 239000006185 dispersion Substances 0.000 description 2
- 239000000843 powder Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000003754 machining Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of image processing, in particular to a filtering processing method of a machine tool cutter image based on a complex industrial environment, which comprises the following steps: collecting a machine tool gray scale map of a machine tool; according to the abnormal gray level change condition of pixel points in the gray level image of the machine tool, dividing the connected domain of the gray level image of the machine tool to obtain a plurality of connected domains to be analyzed of the tool; according to the position distribution and gray level change difference of the connected domain to be analyzed of the cutter in the row-column direction, the horizontal noise degree and the vertical noise degree are obtained; combining the horizontal noise degree and the vertical noise degree to obtain the effectiveness of the cutter information; and selecting and filtering the connected domain to be analyzed of the cutter according to the effectiveness of the cutter information. The invention reduces the noise content of the effective area in the image, reduces the loss of cutter information of the effective area, and improves the filtering efficiency.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a filtering processing method of a machine tool image based on a complex industrial environment.
Background
In modern industrial production, ensuring the detailed features of machine tools plays a critical role in maintaining the machining quality; the existing method is to shoot images of machine tool cutters through an industrial camera, express detailed characteristics of the machine tool cutters through an image processing technology and realize intelligent detection of the machine tool cutters; in a complex industrial environment, the image of a machine tool cutter can be interfered by a certain amount of noise due to the influence of a certain amount of electromagnetic interference and other interference factors such as dust and the like on processing equipment, so that the intelligent detection efficiency of the machine tool cutter is reduced; therefore, in order to improve the intelligent detection efficiency of the machine tool, it is necessary to perform filtering denoising processing on the machine tool image.
In the existing method, an image filtering algorithm is generally utilized to denoise the machine tool image, and noise amounts contained in different areas in corresponding images are different due to uneven distribution of interfering substances such as chips, powder and the like on the surface of the machine tool; the traditional image filtering algorithm only carries out denoising treatment on the whole image, and does not consider the situation of different noise amounts caused by the distribution situation of interfering substances on the surface of a tool of a machine tool, so that a great amount of noise is still reserved on the filtered image, and the filtering efficiency is reduced.
Disclosure of Invention
The invention provides a filtering processing method of a machine tool cutter image based on a complex industrial environment, which aims to solve the existing problems: the traditional image filtering algorithm only carries out denoising treatment on the whole image, and does not consider the situation of different noise amounts caused by the distribution situation of interfering substances on the surface of a machine tool, so that a great amount of noise is still reserved in the filtered image.
The filtering processing method of the machine tool cutter image based on the complex industrial environment adopts the following technical scheme:
The method comprises the following steps:
collecting a machine tool gray scale map of a machine tool;
According to the abnormal gray level change condition of pixel points in the gray level image of the machine tool, dividing the connected domain of the gray level image of the machine tool to obtain a plurality of connected domains to be analyzed of the tool;
According to the position distribution and gray level change difference of the connected domains to be analyzed of the cutters in the row and column directions, horizontal noise degree and vertical noise degree of each connected domain to be analyzed of the cutters are obtained; combining the horizontal noise degree and the vertical noise degree of each cutter communicating region to be analyzed to obtain the cutter information validity of each cutter communicating region to be analyzed;
And screening the connected domain to be analyzed of the cutter according to the cutter information validity of each connected domain to be analyzed of the cutter, obtaining a core filtering region and filtering.
Preferably, the method for dividing the connected domain of the gray level map of the machine tool according to the abnormal gray level change condition of the pixel points in the gray level map of the machine tool to obtain a plurality of connected domains to be analyzed of the tool includes the following specific steps:
For any machine tool gray scale map, clustering the number of pixel points of all gray scale values in the machine tool gray scale map to obtain a plurality of clusters; taking the maximum value of the pixel number of all gray values in all clusters as the numerator of a sample set, taking the minimum value of the pixel number of all gray values in all clusters as the denominator of the sample set, recording the ratio of the numerator of the sample set to the denominator of the sample set as the initial number of the sample set, and recording the integer obtained by rounding the initial number of the sample set as the number of the adaptive sample set ；
Based on adaptive sample set numberThe gray pixel point extraction number of each gray value in each cluster is obtained according to the distribution ratio condition of the number of the pixel points contained in each cluster;
Screening a plurality of abnormal cutter pixel points from all the pixel points according to the extraction number of gray pixel points of each gray value in each cluster;
Marking the average value of the gray values of all abnormal cutter pixels as an initial gray threshold value, and marking the integer after rounding the initial gray threshold value as an adaptive gray value; dividing the machine tool gray level map by taking the self-adaptive gray level value as a dividing threshold value to obtain a machine tool dividing map; and acquiring all the connected domains in the tool segmentation graph of the machine tool, and marking each connected domain as a connected domain to be analyzed of the tool.
Preferably, the method is based on the number of adaptive sample setsThe method for obtaining the extraction number of gray pixel points of each gray value in each cluster according to the distribution ratio condition of the number of pixel points contained in each cluster comprises the following specific steps:
for any cluster, the number of pixel points of all gray values in the cluster and the number of adaptive sample sets are calculated As the initial extraction number of the cluster-like pixel points of the cluster; the initial extraction number of the cluster-like pixel points is rounded to be an integer, and the integer is recorded as the extraction number of the cluster-like pixel points of the cluster; marking any gray value in the cluster as a target gray value, and taking the ratio of the number of pixels of the target gray value to the number of pixels of all gray values in the cluster as an extraction proportionality coefficient of the target gray value; the product of the self-adaptive extraction number of the cluster and the extraction proportion coefficient of the target gray value is recorded as the initial extraction number of gray pixel points of the target gray value; and (3) rounding the initial extraction number of the gray pixel points of the target gray value to form an integer, and recording the integer as the extraction number of the gray pixel points of the target gray value.
Preferably, the method includes the steps of:
Number of sample sets to be adapted As a random extraction total number; the random extraction number of the optional one time is recorded as the random extraction number of the sample, the gray value of the optional one in the optional one cluster is recorded as the marked gray value, and the gray pixel extraction number of the marked gray value is recorded as/>Randomly extracting/>, under the random sampling times of the samples, from the gray level diagram of the tool of the machine toolAfter the number of pixels with the same gray value as the marked gray value are extracted from the gray pixel corresponding to the gray value in all clusters, the set formed by all the pixels extracted randomly at each time is marked as a sample set of the pixels to be analyzed of the cutter; performing isolated forest detection on all the pixels in the pixel sample set to be analyzed of each cutter to obtain the abnormal score of each pixel; presetting an abnormality score threshold/>The anomaly score is greater than/>Is marked as an abnormal pixel point of the cutter.
Preferably, the method for obtaining the horizontal noise degree and the vertical noise degree of the connected domain to be analyzed of each cutter according to the position distribution and the gray level change difference of the connected domain in the row-column direction of the connected domain to be analyzed of the cutter includes the following specific steps:
For a machine tool cutting graph of any machine tool gray graph, marking an optional one tool to-be-analyzed connected domain as a target tool to-be-analyzed connected domain, and marking each tool to-be-analyzed connected domain except the target tool to-be-analyzed connected domain as a to-be-compared tool to-be-analyzed connected domain of the target tool to-be-analyzed connected domain;
According to the area distribution of the target cutter to-be-analyzed connected domains in the row direction and the column direction respectively, screening a plurality of same-row cutter to-be-analyzed connected domains and a plurality of same-column cutter to-be-analyzed connected domains from all connected domains;
Obtaining left and right communication distance contrast values of the target cutter to-be-analyzed communication domain according to the distance ratio between the target cutter to-be-analyzed communication domain and the left and right side same-row cutter to-be-analyzed communication domain; obtaining the upper and lower communication distance comparison value of the target cutter to-be-analyzed communication domain according to the distance ratio between the target cutter to-be-analyzed communication domain and the upper and lower cutter to-be-analyzed communication domains in the same row;
according to the distribution distance change conditions of adjacent same-row cutter to-be-analyzed connected domains and adjacent same-row cutter to-be-analyzed connected domains of the target cutter to-be-analyzed connected domains in the row-column direction, obtaining adjacent connected row distance fluctuation values and adjacent connected column distance fluctuation values of the target cutter to-be-analyzed connected domains;
obtaining a horizontal noise factor and a vertical noise factor of the communicating region to be analyzed of the target cutter according to the left-right communicating distance comparison value, the adjacent communicating line distance fluctuation value, the up-down communicating distance comparison value and the adjacent communicating column distance fluctuation value of the communicating region to be analyzed of the target cutter;
Obtaining horizontal noise factors of all the connected domains to be analyzed of all the cutters, carrying out linear normalization on all the horizontal noise factors, and marking each normalized horizontal noise factor as a horizontal noise degree; and acquiring vertical noise factors of all the connected domains to be analyzed of the cutter, carrying out linear normalization on all the vertical noise factors, and recording each normalized vertical noise factor as a vertical noise degree.
Preferably, the method for screening the connected domain to be analyzed of the same row of cutters and the connected domain to be analyzed of the same column of cutters from all the connected domains according to the area distribution of the connected domain to be analyzed of the target cutters in the row and column directions of other connected domains to be analyzed of the comparative cutters includes the following specific steps:
according to the number of image lines occupied by the connected domain to be analyzed of the target cutter, the number of the connected domain lines is determined in a self-adaptive mode, and a plurality of connected domains to be analyzed of the same-line cutters which accord with the number of the connected domain lines are screened out in the line direction;
according to the number of image columns occupied by the connected domain to be analyzed of the target cutter, the number of the connected domain columns is determined in a self-adaptive mode, and a plurality of connected domains to be analyzed of the same-column cutter, which accord with the number of the connected domain columns, are screened out in the column direction.
Preferably, the method includes the specific steps of:
For any two adjacent same-row cutter to-be-analyzed connected domains of the target cutter to-be-analyzed connected domains, taking the average value of the connected distances between the two same-row cutter to-be-analyzed connected domains and the target cutter to-be-analyzed connected domain as a local row connected distance value of the two same-row cutter to-be-analyzed connected domains; obtaining local row communication distance values of all arbitrary two adjacent same-row cutter to-be-analyzed communication domains of the target cutter to-be-analyzed communication domains, and marking the average value of all the local row communication distance values as the adjacent communication distance fluctuation value of the target cutter to-be-analyzed communication domains;
For any two adjacent same-row tool to-be-analyzed connected domains of the target tool to-be-analyzed connected domains, taking the average value of the connected distances between the two same-row tool to-be-analyzed connected domains and the target tool to-be-analyzed connected domains as a local row connected distance value of the two same-row tool to-be-analyzed connected domains; and acquiring local column communication distance values of all any two adjacent same-column tool to-be-analyzed communication domains of the target tool to-be-analyzed communication domains, and marking the average value of all the local column communication distance values as an adjacent column communication distance fluctuation value of the target tool to-be-analyzed communication domain.
Preferably, the horizontal noise factor and the vertical noise factor of the communicating domain to be analyzed of the target cutter are obtained according to the left-right communicating distance contrast value, the adjacent communicating line distance fluctuation value, the upper-lower communicating distance contrast value and the adjacent communicating column distance fluctuation value of the communicating domain to be analyzed of the target cutter, and the specific method comprises the following steps:
In the method, in the process of the invention, A horizontal noise factor representing a connected domain to be analyzed of the target cutter; /(I)Representing the number of all connected domains to be analyzed of the same-row cutters of the connected domain to be analyzed of the target cutter; /(I)Representing the left and right communication distance contrast value of the communication domain to be analyzed of the target cutter; /(I)Representing adjacent passing distance fluctuation values of a communicating region to be analyzed of a target cutter; /(I)Representing a preset denominator super parameter; /(I)Representing the number of all pixel points in a communication domain to be analyzed of a target cutter; /(I)Represents the/>, of the connected domain to be analyzed of the target toolThe number of all pixel points in the connected domain to be analyzed by the same-row cutters; /(I)The representation takes absolute value.
Preferably, the method for obtaining the validity of the cutter information of each cutter to be analyzed connected domain includes the following specific steps:
Presetting a horizontal noise weight And a vertical noise weight/>Said/>; The horizontal noise degree of the communicating domain to be analyzed of the target cutter is calculated by/>Taking the product of the vertical noise degree and/>, of the connected domain to be analyzed of the target tool, as the effectiveness of the tool horizontal informationTaking the product of the tool vertical information validity as the tool information validity of the connected domain to be analyzed of the target tool, and taking the addition result of the tool horizontal information validity and the tool vertical information validity as the tool information validity of the connected domain to be analyzed of the target tool.
Preferably, the method includes screening the connected domain to be analyzed according to the validity of the cutter information of each connected domain to be analyzed, obtaining a core filtering region, and filtering, including the following specific steps:
Presetting a classification quantity for any machine tool cutter gray scale map Taking absolute values of differences of tool information validity between connected domains to be analyzed of different tools as distance measures, and according to the distance measures and/>Clustering all connected domains to be analyzed of the cutters through a k-means clustering algorithm to obtain a plurality of clusters; marking the mean value of the cutter information effectiveness of all cutter connected domains to be analyzed in each cluster as a core mean value to be selected, and marking the cluster with the largest core mean value to be selected as a core filtering cluster;
The image area occupied by each cutter to-be-analyzed connected domain in the core filtering cluster in the machine tool cutter segmentation map is marked as a comparison filtering area, and the image area with the same position as the comparison filtering area in the machine tool cutter gray map is marked as a core filtering area; and (5) marking the gray scale map of the machine tool after denoising all the core filtering areas as a filtering machine tool image.
The technical scheme of the invention has the beneficial effects that: according to the method, a plurality of connected domains to be analyzed of the tool are obtained through abnormal gray level change conditions of pixel points in a gray level image of the tool of the machine tool; according to the position distribution and gray level change difference of the connected domain of the cutter to be analyzed in the row direction, the horizontal noise degree and the vertical noise degree of the connected domain of the cutter to be analyzed are obtained, and then the cutter information effectiveness of the connected domain of the cutter to be analyzed is obtained; selecting a connected domain to be analyzed of the cutter according to the effectiveness of the cutter information, and filtering; firstly, a plurality of cutter to-be-analyzed connected domains are obtained through abnormal gray level change conditions of pixel points in a gray level image of a cutter of a machine tool, the cutter to-be-analyzed connected domains are used for describing a cutter effective area needing denoising, and the cutter effective area needing denoising is screened out through preliminary analysis of abnormal pixel point distribution conditions, so that filtering cost is reduced; then according to the position distribution and gray level change difference of the connected domain of the cutter to be analyzed in the row and column directions, obtaining the horizontal noise degree and the vertical noise degree of the connected domain of the cutter to be analyzed, wherein the horizontal noise degree and the vertical noise degree are respectively used for describing the degree of denoising of the connected domain of the cutter to be analyzed in the row and column directions, so that the profit difference of the cutter after denoising in different directions is more obvious; then according to the horizontal noise degree and the vertical noise degree, obtaining the cutter information validity of the connected domain to be analyzed of the cutter, wherein the cutter information validity is used for describing the effective cutter information content contained in the whole connected domain to be analyzed of the cutter, and reducing the loss of effective cutter information after denoising; finally, according to the cutter information validity of each cutter connected domain to be analyzed, selecting and filtering the cutter connected domain to be analyzed; the filtering cost is reduced, the noise content of an effective area in an image is reduced, meanwhile, the loss of cutter information of the effective area is reduced, and the filtering efficiency is improved.
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 steps of a method for filtering a machine tool image based on a complex industrial environment of the present invention;
FIG. 2 is a schematic illustration of a gray scale map of a machine tool bit of the present invention;
FIG. 3 is a schematic view of a machine tool cutter split map of the present invention;
FIG. 4 is a schematic diagram of a filtered tool map of the machine tool after denoising according to the conventional image filtering algorithm of the present invention;
fig. 5 is a schematic diagram of a filtered machine tool cutter map after denoising with the improved image filtering algorithm 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 detailed description is given below of the filtering processing method of the machine tool cutter image based on the complex industrial environment according to the invention, which is provided by the invention, with reference to the attached 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 filtering processing method of the machine tool cutter image based on the complex industrial environment provided by the invention with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of a filtering processing method for a machine tool cutter image based on a complex industrial environment according to an embodiment of the present invention is shown, where the method includes the following steps:
step S001: and collecting a machine tool gray scale map of the machine tool.
Specifically, firstly, a gray level diagram of a machine tool cutter needs to be acquired, and the specific process is as follows: and shooting machine tool cutter images by using an industrial camera, and carrying out gray-scale treatment on each machine tool cutter image to obtain a machine tool cutter gray-scale image. The graying process is a known technique, and the description of this embodiment is omitted. Referring to fig. 2, a schematic diagram of a gray scale map of a machine tool bit is shown.
So far, the gray level map of the machine tool cutter is obtained through the method.
Step S002: and according to the abnormal gray level change condition of the pixel points in the gray level map of the machine tool, dividing the connected domain of the gray level map of the machine tool to obtain a plurality of connected domains to be analyzed of the tool.
It should be noted that when the conventional image filtering algorithm is used to denoise the machine tool image, the noise amounts contained in different areas in the corresponding image are different due to uneven distribution of the interfering substances such as chips, powder and the like on the surface of the machine tool; the traditional image filtering algorithm only carries out denoising treatment on the whole image, and does not consider the situation of different noise amounts caused by the distribution situation of interfering substances on the surface of a machine tool, so that a great amount of noise is still reserved in the filtered image, and the filtering efficiency is reduced; therefore, in order to improve the filtering efficiency, the degree of noise quantity contained in different areas can be determined by analyzing the distribution condition of interference substances in different areas in a machine tool gray scale map, and the traditional image filtering algorithm is used for denoising.
It should be further noted that before analyzing the distribution of the interfering substances in different areas in the gray scale of the tool of the machine tool, the different areas need to be divided; the partial areas are divided differently, and the influence on the subsequent filtering efficiency is also different, so that the adaptive division of the areas is required; in order to ensure that the efficiency of a filtering result is not greatly interfered, the existing method generally divides a machine tool gray level diagram through determining a gray level threshold value to obtain a corresponding region to be analyzed; however, a certain amount of pixel points under abnormal conditions exist in the actual gray level diagram of the machine tool, and the pixel points are usually pixel points in other tiny parts or background areas except the machine tool, so that proper selection of gray level thresholds is greatly interfered, and therefore the influence condition of the pixel points under abnormal conditions on gray level threshold selection needs to be analyzed, proper segmentation thresholds are determined, and the gray level diagram of the machine tool is segmented to obtain corresponding connected domain areas for subsequent analysis and processing.
Preferably, in one embodiment of the present invention, according to the abnormal condition of gray level change of a pixel point in a gray level map of a tool of a machine tool, the gray level map of the tool of the machine tool is divided into connected domains to obtain a plurality of connected domains to be analyzed of the tool, including the specific method that:
taking a gray level diagram of any machine tool as an example, presetting the number of cluster clusters Wherein the present embodiment usesTo describe the example, the present embodiment is not particularly limited, wherein/>Depending on the particular implementation; taking the absolute value of the difference value of the pixel point numbers among different gray values as a distance measure, and according to the distance measure and the cluster number/>Clustering the number of pixel points of all gray values in the gray map of the machine tool to obtain a plurality of clusters; taking the maximum value of the pixel number of all gray values in all clusters as the numerator of a sample set, taking the minimum value of the pixel number of all gray values in all clusters as the denominator of the sample set, recording the ratio of the numerator of the sample set to the denominator of the sample set as the initial number of the sample set, and recording the integer obtained by rounding the initial number of the sample set as the number/>. Wherein each cluster contains a plurality of gray values; in addition, the clustering process according to the distance measurement and the number of clusters is a well-known content of the k-means clustering algorithm, and the embodiment is not described in detail.
Further, taking any cluster as an example, the number of pixel points of all gray values in the cluster and the number of adaptive sample sets are calculatedAs the initial extraction number of the cluster-like pixel points of the cluster; the initial extraction number of the cluster-like pixel points is rounded to be an integer, and the integer is recorded as the extraction number of the cluster-like pixel points of the cluster; marking any gray value in the cluster as a target gray value, and taking the ratio of the number of pixels of the target gray value to the number of pixels of all gray values in the cluster as an extraction proportionality coefficient of the target gray value; the product of the self-adaptive extraction number of the cluster and the extraction proportion coefficient of the target gray value is recorded as the initial extraction number of gray pixel points of the target gray value; the integer after rounding the initial extraction number of the gray pixel points of the target gray value is marked as the extraction number of the gray pixel points of the target gray value; acquiring the extraction number of gray pixel points of each gray value in the cluster; and acquiring the extraction number of gray pixel points of each gray value in each cluster.
Further, the number of sample sets is adaptedAs a random extraction total number; the random extraction number of the optional one time is recorded as the random extraction number of the sample, the gray value of the optional one in the optional one cluster is recorded as the marked gray value, and the gray pixel extraction number of the marked gray value is recorded as/>Randomly extracting/>, under the random sampling times of the samples, from the gray level diagram of the tool of the machine toolAnd after the number of the pixels with the same gray value as the marked gray value are extracted from the gray pixel corresponding to the gray value in all the clusters, marking the set formed by all the pixels randomly extracted each time as a sample set of the pixels to be analyzed of the cutter. And performing isolated forest detection on all the pixels in the pixel sample set to be analyzed of each cutter to obtain the anomaly score of each pixel. Presetting an abnormality score threshold/>Wherein the present embodiment usesTo describe the example, the present embodiment is not particularly limited, wherein/>Depending on the particular implementation; score of abnormality greater than/>Is marked as an abnormal pixel point of the cutter. Wherein the pixel points contained between the pixel point sample sets to be analyzed of different cutters are not identical; in addition, the process of randomly extracting the pixel points is a well-known content of an isolated forest algorithm, and the embodiment is not repeated.
Further, the average value of the gray values of all abnormal cutter pixel points is recorded as an initial gray threshold value, and the integer obtained by rounding the initial gray threshold value is recorded as an adaptive gray value; taking the self-adaptive gray value as a segmentation threshold value, and segmenting the machine tool gray map according to the segmentation threshold value to obtain a machine tool segmentation map; and acquiring all the connected domains in the tool segmentation graph of the machine tool, and marking each connected domain as a connected domain to be analyzed of the tool. The process of dividing the image according to the division threshold is a well-known content of the division algorithm of the oxford threshold, and this embodiment will not be described in detail. Referring to fig. 3, a schematic diagram of a machine tool bit split map is shown.
So far, the method is used for obtaining a plurality of connected domains to be analyzed of the cutters.
Step S003: according to the position distribution and gray level change difference of the connected domains to be analyzed of the cutters in the row and column directions, horizontal noise degree and vertical noise degree of each connected domain to be analyzed of the cutters are obtained; and combining the horizontal noise degree and the vertical noise degree of each cutter communicating region to be analyzed to obtain the cutter information validity of each cutter communicating region to be analyzed.
The method is characterized in that a to-be-analyzed connected domain of the tool obtained through the self-adaptive segmentation threshold is a core operation region of the tool of the machine tool; the core operation area of the machine tool cutter can be vertically placed in daily work and then is operated; the more complete the distribution information content of the tool area of the machine tool in the vertical direction is compared with the distribution information content in the horizontal direction in the communication area to be analyzed of the corresponding tool; meanwhile, when the machine tool cutter is used for processing a metal object, the metal object can be horizontally placed below the cutter head of the machine tool cutter, and when the filtering treatment is carried out on the machine tool cutter, the area except the machine tool cutter is not important, so that the important filtering is not needed for the metal object; therefore, it is necessary to determine the horizontal noise factor and the vertical noise factor by analyzing the distribution distance of the pixels in the row and column directions in the connected domain to be analyzed of the tool, and then combine the two to obtain the validity of the tool information for subsequent filtering processing.
Preferably, in one embodiment of the present invention, according to the position distribution and the gray level variation difference of the connected domain of the to-be-analyzed connected domain of the cutter in the row and column directions, the horizontal noise level and the vertical noise level of each connected domain of the to-be-analyzed cutter are obtained, including the following specific methods:
taking a machine tool cutting tool dividing diagram of any machine tool cutting tool gray level diagram as an example, marking each row occupied by each cutting tool communicating domain to be analyzed in the machine tool cutting tool dividing diagram as an image communicating domain to be analyzed of each cutting tool communicating domain; marking each row of the connected domain to be analyzed of each cutter in the cutter segmentation diagram of the machine tool as an image connected row of the connected domain to be analyzed of each cutter; taking Euclidean distance of the mass center between the communicating areas to be analyzed of any two cutters as the communicating distance of the communicating areas to be analyzed of the two cutters; and marking the optional one cutter to-be-analyzed connected domain as a target cutter to-be-analyzed connected domain, and marking each cutter to-be-analyzed connected domain except the target cutter to-be-analyzed connected domain as a cutter to-be-analyzed connected domain to be compared of the target cutter to-be-analyzed connected domain.
Further, taking any one of the to-be-compared cutter to be analyzed connected domain as an example, if each image connected domain of the target cutter to be analyzed connected domain is contained in the image connected domain of the to-be-compared cutter to be analyzed connected domain, marking the to-be-compared cutter to be analyzed connected domain as a first comparison line cutter to be analyzed connected domain of the target cutter to be analyzed connected domain; if each image passing channel of the comparison tool communicating domain to be analyzed is contained in the image passing channel of the target tool communicating domain to be analyzed, marking the comparison tool communicating domain to be analyzed as a second comparison line tool communicating domain of the target tool communicating domain to be analyzed; acquiring all first comparison line cutter to-be-analyzed connected domains and all second comparison line cutter to-be-analyzed connected domains of the target cutter to-be-analyzed connected domains; and marking each first comparison line cutter to-be-analyzed connected domain and each second comparison line cutter to-be-analyzed connected domain of the target cutter to-be-analyzed connected domain as the same line cutter to-be-analyzed connected domain of the target cutter to-be-analyzed connected domain. In this embodiment, the connected domains to be analyzed of the same-row cutter of the connected domains to be analyzed of the default target cutter are ordered according to the sequence from left to right.
Further, taking any one of the to-be-compared cutter to be analyzed connected domain as an example, if each image connected column of the target cutter to be analyzed connected domain is contained in the image connected column of the to-be-compared cutter to be analyzed connected domain, marking the to-be-compared cutter to be analyzed connected domain as a first comparison column cutter to be analyzed connected domain of the target cutter to be analyzed connected domain; if each image communication column of the comparison tool communication domain to be analyzed is contained in the image communication column of the target tool communication domain to be analyzed, marking the comparison tool communication domain to be analyzed as a second comparison column tool communication domain of the target tool communication domain to be analyzed; acquiring all first contrast row cutter to-be-analyzed connected domains and all second contrast row cutter to-be-analyzed connected domains of the target cutter to-be-analyzed connected domains; and marking each first comparison array cutter to-be-analyzed connected domain and each second comparison array cutter to-be-analyzed connected domain of the target cutter to-be-analyzed connected domain as the same array cutter to-be-analyzed connected domain of the target cutter to-be-analyzed connected domain. The connected domains to be analyzed of the same-row cutter of the connected domains to be analyzed of the default target cutter are ordered according to the sequence from top to bottom.
Further, the accumulated sum of the communication distances between all the same-row tool to-be-analyzed communication domains on the left side of the target tool to-be-analyzed communication domain and the target tool to-be-analyzed communication domain is recorded as a left communication distance value of the target tool to-be-analyzed communication domain; the accumulated sum of the communication distances between all the same-row cutter to-be-analyzed communication domains on the right side of the target cutter to-be-analyzed communication domain and the target cutter to-be-analyzed communication domain is recorded as a right communication distance value of the target cutter to-be-analyzed communication domain; and (3) marking the average value of the left communication distance value of the communication domain to be analyzed of the target cutter and the right communication distance value of the communication domain to be analyzed of the target cutter as the left and right communication distance comparison value of the communication domain to be analyzed of the target cutter. Specifically, if the left side of the communicating region to be analyzed of the target cutter does not have the communicating region to be analyzed of the same row of cutters, presetting the left communicating distance value of the communicating region to be analyzed of the target cutter as1, and acquiring a left communicating distance comparison value and a right communicating distance comparison value; if the communicating domain to be analyzed of the target cutter does not exist on the right side of the communicating domain to be analyzed of the target cutter, presetting the right communicating distance value of the communicating domain to be analyzed of the target cutter to be 1, and acquiring the left and right communicating distance comparison value.
Further, the accumulated sum of the communication distances between all the same-row tool to-be-analyzed communication domains above the target tool to-be-analyzed communication domain and the target tool to-be-analyzed communication domain is recorded as an upper communication distance value of the target tool to-be-analyzed communication domain; the accumulated sum of the communication distances between all the same-row cutter to-be-analyzed communication domains below the target cutter to-be-analyzed communication domain and the target cutter to-be-analyzed communication domain is recorded as a lower communication distance value of the target cutter to-be-analyzed communication domain; and (3) marking the average value of the upper communication distance value of the communication domain to be analyzed of the target cutter and the lower communication distance value of the communication domain to be analyzed of the target cutter as the upper and lower communication distance comparison value of the communication domain to be analyzed of the target cutter. Specifically, if the same row of tool to-be-analyzed connected domains do not exist above the target tool to-be-analyzed connected domains, presetting the upper connected distance value of the target tool to-be-analyzed connected domains to be 1, and acquiring the upper and lower connected distance comparison values; if the connected domain to be analyzed of the target cutter does not exist below the connected domain to be analyzed of the target cutter, presetting the lower connected distance value of the connected domain to be analyzed of the target cutter to be 1, and acquiring the upper and lower connected distance comparison value.
Further, taking any two adjacent connected-line tool to-be-analyzed connected-domain of the target tool to-be-analyzed connected-domain as an example, taking the average value of the connected distances between the two connected-line tool to-be-analyzed connected-domains and the target tool to-be-analyzed connected-domain as the local line connected-distance value of the two connected-line tool to-be-analyzed connected-domains; and acquiring local row communication distance values of all any two adjacent same-row cutter to-be-analyzed communication domains of the target cutter to-be-analyzed communication domains, and marking the average value of all the local row communication distance values as the adjacent communication distance fluctuation value of the target cutter to-be-analyzed communication domains.
Further, taking any two adjacent same-row cutter to-be-analyzed connected domains of the target cutter to-be-analyzed connected domains as an example, taking the average value of the connected distances between the two same-row cutter to-be-analyzed connected domains and the target cutter to-be-analyzed connected domains as a local row connected distance value of the two same-row cutter to-be-analyzed connected domains; and acquiring local column communication distance values of all any two adjacent same-column tool to-be-analyzed communication domains of the target tool to-be-analyzed communication domains, and marking the average value of all the local column communication distance values as an adjacent column communication distance fluctuation value of the target tool to-be-analyzed communication domain.
Further, according to the left and right communication distance contrast value and the adjacent communication line distance fluctuation value of the communication domain to be analyzed of the target cutter, the horizontal noise factor of the communication domain to be analyzed of the target cutter is obtained. As an example, the horizontal noise factor of the connected domain to be analyzed of the target tool may be calculated by the following formula:
In the method, in the process of the invention, A horizontal noise factor representing a connected domain to be analyzed of the target cutter; /(I)Representing the number of all connected domains to be analyzed of the same-row cutters of the connected domain to be analyzed of the target cutter; /(I)Representing the left and right communication distance contrast value of the communication domain to be analyzed of the target cutter; /(I)Representing adjacent passing distance fluctuation values of a communicating region to be analyzed of a target cutter; /(I)Representing a predetermined denominator hyper-parameter, the present embodiment is predetermined/>For preventing denominator from being 0; /(I)Representing the number of all pixel points in a communication domain to be analyzed of a target cutter; /(I)Represents the/>, of the connected domain to be analyzed of the target toolThe number of all pixel points in the connected domain to be analyzed by the same-row cutters; /(I)The representation takes absolute value. And obtaining horizontal noise factors of the connected domain to be analyzed of all the cutters, carrying out linear normalization on all the horizontal noise factors, and recording each normalized horizontal noise factor as the horizontal noise degree.
It is to be noted that,Representing the dispersion degree of the analysis of the tool to-be-analyzed connected domain of the target tool in the horizontal direction; if the horizontal noise factor of the communication domain to be analyzed of the target cutter is larger, the probability that the communication domain to be analyzed of the target cutter belongs to the cutter region of the machine tool in the horizontal direction is higher, and the noise contained in the communication domain to be analyzed of the target cutter in the horizontal direction is reflected to be removed.
Further, according to the up-down communication distance contrast value and the adjacent communication column distance fluctuation value of the communication domain to be analyzed of the target cutter, the vertical noise factor of the communication domain to be analyzed of the target cutter is obtained. As an example, the vertical noise factor of the connected domain to be analyzed of the target tool may be calculated by the following formula:
In the method, in the process of the invention, Representing the vertical noise factor of the connected domain to be analyzed of the target cutter; /(I)Representing the number of all the connected domains to be analyzed of the same-row tools of the connected domain to be analyzed of the target tool; /(I)Representing the up-down communication distance contrast value of the communication domain to be analyzed of the target cutter; /(I)Representing the adjacent connected column distance fluctuation value of the connected domain to be analyzed of the target cutter; /(I)Representing a preset first denominator hyper-parameter, preset/>, this embodimentFor preventing denominator from being 0; /(I)Representing the number of all pixel points in a communication domain to be analyzed of a target cutter; /(I)Represents the/>, of the connected domain to be analyzed of the target toolThe number of all pixel points in the connected domain to be analyzed by the same-row cutters; /(I)The representation takes absolute value. And acquiring vertical noise factors of all the connected domains to be analyzed of the cutter, carrying out linear normalization on all the vertical noise factors, and recording each normalized vertical noise factor as a vertical noise degree.
It is to be noted that,Representing the dispersion degree of the analysis of the tool to-be-analyzed connected domain of the target tool in the vertical direction; if the vertical noise factor of the communication domain to be analyzed of the target cutter is larger, the probability that the communication domain to be analyzed of the target cutter belongs to the cutter region of the machine tool in the vertical direction is higher, and the noise contained in the communication domain to be analyzed of the target cutter in the vertical direction is reflected to be removed.
The importance of the image information contained in the connected domain to be analyzed in the vertical direction of the cutter is larger than that in the horizontal direction; therefore, in the process of combining the horizontal noise factor and the vertical noise factor to obtain the tool information validity, the weight assigned to the horizontal noise factor needs to be smaller than the weight assigned to the vertical noise factor.
Preferably, in one embodiment of the present invention, combining the horizontal noise level and the vertical noise level of each connected domain to be analyzed by each tool to obtain the validity of the tool information of each connected domain to be analyzed by each tool, including the specific method that:
Presetting a horizontal noise weight And a vertical noise weight/>Wherein the present embodiment usesAnd/>The present embodiment is not specifically limited, and will be described by way of exampleDepending on the particular implementation; the horizontal noise degree of the communicating domain to be analyzed of the target cutter is calculated by/>Taking the product of the vertical noise degree and/>, of the connected domain to be analyzed of the target tool, as the effectiveness of the tool horizontal informationTaking the product of the tool vertical information validity as the tool information validity of the connected domain to be analyzed of the target tool, and taking the addition result of the tool horizontal information validity and the tool vertical information validity as the tool information validity of the connected domain to be analyzed of the target tool. And acquiring the cutter information validity of each cutter connected domain to be analyzed.
So far, the cutter information validity of each cutter connected domain to be analyzed is obtained through the method.
Step S004: and screening the connected domain to be analyzed of the cutter according to the cutter information validity of each connected domain to be analyzed of the cutter, obtaining a core filtering region and filtering.
Preferably, in one embodiment of the present invention, according to the validity of the tool information of each connected domain to be analyzed by the tool, the method for selectively filtering the connected domain to be analyzed by the tool includes:
Taking any machine tool gray scale as an example, presetting a classification number Wherein the present embodiment is described as/>To describe the example, the present embodiment is not particularly limited, wherein/>Depending on the particular implementation; taking absolute values of differences of tool information validity between connected domains to be analyzed of different tools as distance measures, and according to the distance measures and the distance measuresClustering all connected domains to be analyzed of the cutters through a k-means clustering algorithm to obtain a plurality of clusters; and marking the mean value of the cutter information effectiveness of all the cutter connected domains to be analyzed in each cluster as a core mean value to be selected, and marking the cluster with the largest core mean value to be selected as a core filtering cluster.
Further, an image area occupied by each cutter to-be-analyzed connected domain in the core filtering cluster in the machine tool cutter segmentation map is recorded as a comparison filtering area, and an image area with the same position as the comparison filtering area in the machine tool cutter gray scale map is recorded as a core filtering area; and acquiring all the core filtering areas, and marking the gray level map of the machine tool cutter after denoising of all the core filtering areas as a filtering machine tool cutter image. The process of denoising the image is well known in the conventional image filtering algorithm, and this embodiment is not repeated. Referring to fig. 4, a schematic diagram of a filtered tool map of a conventional image filtering algorithm after denoising is shown; referring to fig. 5, a schematic diagram of a filtered machine tool cutter map after denoising with the improved image filtering algorithm is shown.
This embodiment is completed.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the invention, but any modifications, equivalent substitutions, improvements, etc. within the principles of the present invention should be included in the scope of the present invention.
Claims (10)
1. The filtering processing method of the machine tool cutter image based on the complex industrial environment is characterized by comprising the following steps of:
collecting a machine tool gray scale map of a machine tool;
According to the abnormal gray level change condition of pixel points in the gray level image of the machine tool, dividing the connected domain of the gray level image of the machine tool to obtain a plurality of connected domains to be analyzed of the tool;
According to the position distribution and gray level change difference of the connected domains to be analyzed of the cutters in the row and column directions, horizontal noise degree and vertical noise degree of each connected domain to be analyzed of the cutters are obtained; combining the horizontal noise degree and the vertical noise degree of each cutter communicating region to be analyzed to obtain the cutter information validity of each cutter communicating region to be analyzed;
And screening the connected domain to be analyzed of the cutter according to the cutter information validity of each connected domain to be analyzed of the cutter, obtaining a core filtering region and filtering.
2. The filtering processing method of machine tool cutter images based on complex industrial environment according to claim 1, wherein the method for dividing the machine tool cutter gray scale map into connected domains according to the abnormal gray scale change condition of pixel points in the machine tool cutter gray scale map to obtain a plurality of connected domains to be analyzed of the cutter comprises the following specific steps:
For any machine tool gray scale map, clustering the number of pixel points of all gray scale values in the machine tool gray scale map to obtain a plurality of clusters; taking the maximum value of the pixel number of all gray values in all clusters as the numerator of a sample set, taking the minimum value of the pixel number of all gray values in all clusters as the denominator of the sample set, recording the ratio of the numerator of the sample set to the denominator of the sample set as the initial number of the sample set, and recording the integer obtained by rounding the initial number of the sample set as the number of the adaptive sample set ；
Based on adaptive sample set numberThe gray pixel point extraction number of each gray value in each cluster is obtained according to the distribution ratio condition of the number of the pixel points contained in each cluster;
Screening a plurality of abnormal cutter pixel points from all the pixel points according to the extraction number of gray pixel points of each gray value in each cluster;
Marking the average value of the gray values of all abnormal cutter pixels as an initial gray threshold value, and marking the integer after rounding the initial gray threshold value as an adaptive gray value; dividing the machine tool gray level map by taking the self-adaptive gray level value as a dividing threshold value to obtain a machine tool dividing map; and acquiring all the connected domains in the tool segmentation graph of the machine tool, and marking each connected domain as a connected domain to be analyzed of the tool.
3. The method for filtering machine tool images based on complex industrial environment according to claim 2, wherein the adaptive sample set numberThe method for obtaining the extraction number of gray pixel points of each gray value in each cluster according to the distribution ratio condition of the number of pixel points contained in each cluster comprises the following specific steps:
for any cluster, the number of pixel points of all gray values in the cluster and the number of adaptive sample sets are calculated As the initial extraction number of the cluster-like pixel points of the cluster; the initial extraction number of the cluster-like pixel points is rounded to be an integer, and the integer is recorded as the extraction number of the cluster-like pixel points of the cluster; marking any gray value in the cluster as a target gray value, and taking the ratio of the number of pixels of the target gray value to the number of pixels of all gray values in the cluster as an extraction proportionality coefficient of the target gray value; the product of the self-adaptive extraction number of the cluster and the extraction proportion coefficient of the target gray value is recorded as the initial extraction number of gray pixel points of the target gray value; and (3) rounding the initial extraction number of the gray pixel points of the target gray value to form an integer, and recording the integer as the extraction number of the gray pixel points of the target gray value.
4. The filtering processing method of machine tool cutter images based on complex industrial environment according to claim 2, wherein the specific method for screening a plurality of abnormal cutter pixels from all pixels according to the number of gray pixel extraction of each gray value in each cluster comprises the following steps:
Number of sample sets to be adapted As a random extraction total number; the random extraction number of the optional one time is recorded as the random extraction number of the sample, the gray value of the optional one in the optional one cluster is recorded as the marked gray value, and the gray pixel extraction number of the marked gray value is recorded as/>Randomly extracting/>, under the random sampling times of the samples, from the gray level diagram of the tool of the machine toolAfter the number of pixels with the same gray value as the marked gray value are extracted from the gray pixel corresponding to the gray value in all clusters, the set formed by all the pixels extracted randomly at each time is marked as a sample set of the pixels to be analyzed of the cutter; performing isolated forest detection on all the pixels in the pixel sample set to be analyzed of each cutter to obtain the abnormal score of each pixel; presetting an abnormality score threshold/>The anomaly score is greater than/>Is marked as an abnormal pixel point of the cutter.
5. The filtering processing method of machine tool cutter images based on complex industrial environment according to claim 1, wherein the obtaining of the horizontal noise level and the vertical noise level of each cutter to be analyzed connected domain according to the position distribution and the gray level variation difference of the connected domain in the row and column directions of the cutter to be analyzed comprises the following specific steps:
For a machine tool cutting graph of any machine tool gray graph, marking an optional one tool to-be-analyzed connected domain as a target tool to-be-analyzed connected domain, and marking each tool to-be-analyzed connected domain except the target tool to-be-analyzed connected domain as a to-be-compared tool to-be-analyzed connected domain of the target tool to-be-analyzed connected domain;
According to the area distribution of the target cutter to-be-analyzed connected domains in the row direction and the column direction respectively, screening a plurality of same-row cutter to-be-analyzed connected domains and a plurality of same-column cutter to-be-analyzed connected domains from all connected domains;
Obtaining left and right communication distance contrast values of the target cutter to-be-analyzed communication domain according to the distance ratio between the target cutter to-be-analyzed communication domain and the left and right side same-row cutter to-be-analyzed communication domain; obtaining the upper and lower communication distance comparison value of the target cutter to-be-analyzed communication domain according to the distance ratio between the target cutter to-be-analyzed communication domain and the upper and lower cutter to-be-analyzed communication domains in the same row;
according to the distribution distance change conditions of adjacent same-row cutter to-be-analyzed connected domains and adjacent same-row cutter to-be-analyzed connected domains of the target cutter to-be-analyzed connected domains in the row-column direction, obtaining adjacent connected row distance fluctuation values and adjacent connected column distance fluctuation values of the target cutter to-be-analyzed connected domains;
obtaining a horizontal noise factor and a vertical noise factor of the communicating region to be analyzed of the target cutter according to the left-right communicating distance comparison value, the adjacent communicating line distance fluctuation value, the up-down communicating distance comparison value and the adjacent communicating column distance fluctuation value of the communicating region to be analyzed of the target cutter;
Obtaining horizontal noise factors of all the connected domains to be analyzed of all the cutters, carrying out linear normalization on all the horizontal noise factors, and marking each normalized horizontal noise factor as a horizontal noise degree; and acquiring vertical noise factors of all the connected domains to be analyzed of the cutter, carrying out linear normalization on all the vertical noise factors, and recording each normalized vertical noise factor as a vertical noise degree.
6. The filtering processing method of machine tool images based on complex industrial environment according to claim 5, wherein the specific method for screening a plurality of connected tool to be analyzed connected domains and a plurality of connected tool to be analyzed connected domains in the same row from all connected domains according to the area distribution of other connected domains to be analyzed of the target tool to be analyzed in the row and column directions, respectively, comprises the following steps:
according to the number of image lines occupied by the connected domain to be analyzed of the target cutter, the number of the connected domain lines is determined in a self-adaptive mode, and a plurality of connected domains to be analyzed of the same-line cutters which accord with the number of the connected domain lines are screened out in the line direction;
according to the number of image columns occupied by the connected domain to be analyzed of the target cutter, the number of the connected domain columns is determined in a self-adaptive mode, and a plurality of connected domains to be analyzed of the same-column cutter, which accord with the number of the connected domain columns, are screened out in the column direction.
7. The filtering processing method of machine tool images based on complex industrial environment according to claim 5, wherein the obtaining the adjacent connected row distance fluctuation value and the adjacent connected column distance fluctuation value of the target tool to-be-analyzed connected domain according to the distribution distance change condition of the adjacent connected row tool to-be-analyzed connected domain and the adjacent connected column tool to-be-analyzed connected domain of the target tool to-be-analyzed connected domain in the row and column directions respectively comprises the following specific steps:
For any two adjacent same-row cutter to-be-analyzed connected domains of the target cutter to-be-analyzed connected domains, taking the average value of the connected distances between the two same-row cutter to-be-analyzed connected domains and the target cutter to-be-analyzed connected domain as a local row connected distance value of the two same-row cutter to-be-analyzed connected domains; obtaining local row communication distance values of all arbitrary two adjacent same-row cutter to-be-analyzed communication domains of the target cutter to-be-analyzed communication domains, and marking the average value of all the local row communication distance values as the adjacent communication distance fluctuation value of the target cutter to-be-analyzed communication domains;
For any two adjacent same-row tool to-be-analyzed connected domains of the target tool to-be-analyzed connected domains, taking the average value of the connected distances between the two same-row tool to-be-analyzed connected domains and the target tool to-be-analyzed connected domains as a local row connected distance value of the two same-row tool to-be-analyzed connected domains; and acquiring local column communication distance values of all any two adjacent same-column tool to-be-analyzed communication domains of the target tool to-be-analyzed communication domains, and marking the average value of all the local column communication distance values as an adjacent column communication distance fluctuation value of the target tool to-be-analyzed communication domain.
8. The filtering processing method for machine tool images based on complex industrial environments according to claim 5, wherein the obtaining the horizontal noise factor and the vertical noise factor of the communicating domain to be analyzed according to the left-right communicating distance contrast value, the adjacent communicating line distance fluctuation value, the upper-lower communicating distance contrast value and the adjacent communicating column distance fluctuation value of the communicating domain to be analyzed by the target tool comprises the following specific steps:
In the method, in the process of the invention, A horizontal noise factor representing a connected domain to be analyzed of the target cutter; /(I)Representing the number of all connected domains to be analyzed of the same-row cutters of the connected domain to be analyzed of the target cutter; /(I)Representing the left and right communication distance contrast value of the communication domain to be analyzed of the target cutter; /(I)Representing adjacent passing distance fluctuation values of a communicating region to be analyzed of a target cutter; /(I)Representing a preset denominator super parameter; Representing the number of all pixel points in a communication domain to be analyzed of a target cutter; /(I) Represents the/>, of the connected domain to be analyzed of the target toolThe number of all pixel points in the connected domain to be analyzed by the same-row cutters; /(I)The representation takes absolute value.
9. The filtering processing method of machine tool cutter images based on complex industrial environment according to claim 1, wherein the combining of the horizontal noise level and the vertical noise level of each cutter to-be-analyzed connected domain to obtain the cutter information validity of each cutter to-be-analyzed connected domain comprises the following specific steps:
Presetting a horizontal noise weight And a vertical noise weight/>Said/>; The horizontal noise degree of the communicating domain to be analyzed of the target cutter is calculated by/>Taking the product of the vertical noise degree and/>, of the connected domain to be analyzed of the target tool, as the effectiveness of the tool horizontal informationTaking the product of the tool vertical information validity as the tool information validity of the connected domain to be analyzed of the target tool, and taking the addition result of the tool horizontal information validity and the tool vertical information validity as the tool information validity of the connected domain to be analyzed of the target tool.
10. The filtering processing method of machine tool images based on complex industrial environment according to claim 1, wherein the specific method for screening the connected domain to be analyzed according to the tool information validity of each connected domain to be analyzed, obtaining a core filtering region and filtering comprises the following steps:
Presetting a classification quantity for any machine tool cutter gray scale map Taking absolute values of differences of tool information validity between connected domains to be analyzed of different tools as distance measures, and according to the distance measures and/>Clustering all connected domains to be analyzed of the cutters through a k-means clustering algorithm to obtain a plurality of clusters; marking the mean value of the cutter information effectiveness of all cutter connected domains to be analyzed in each cluster as a core mean value to be selected, and marking the cluster with the largest core mean value to be selected as a core filtering cluster;
The image area occupied by each cutter to-be-analyzed connected domain in the core filtering cluster in the machine tool cutter segmentation map is marked as a comparison filtering area, and the image area with the same position as the comparison filtering area in the machine tool cutter gray map is marked as a core filtering area; and (5) marking the gray scale map of the machine tool after denoising all the core filtering areas as a filtering machine tool image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410572336.9A CN118154458B (en) | 2024-05-10 | 2024-05-10 | Filtering processing method of machine tool cutter image based on complex industrial environment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410572336.9A CN118154458B (en) | 2024-05-10 | 2024-05-10 | Filtering processing method of machine tool cutter image based on complex industrial environment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118154458A true CN118154458A (en) | 2024-06-07 |
| CN118154458B CN118154458B (en) | 2024-08-20 |
Family
ID=91291021
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410572336.9A Active CN118154458B (en) | 2024-05-10 | 2024-05-10 | Filtering processing method of machine tool cutter image based on complex industrial environment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118154458B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020198997A1 (en) * | 2019-03-29 | 2020-10-08 | 深圳市大疆创新科技有限公司 | Connected domain processing method, data processing device and computer readable storage medium |
| CN114998112A (en) * | 2022-04-22 | 2022-09-02 | 南通悦福软件有限公司 | Image denoising method and system based on adaptive frequency domain filtering |
| CN116703910A (en) * | 2023-08-07 | 2023-09-05 | 威海丰荟建筑工业科技有限公司 | Intelligent detection method for quality of concrete prefabricated bottom plate |
| CN117237245A (en) * | 2023-11-16 | 2023-12-15 | 湖南云箭智能科技有限公司 | Industrial material quality monitoring method based on artificial intelligence and Internet of things |
| CN117809379A (en) * | 2024-02-29 | 2024-04-02 | 深圳市翔飞科技股份有限公司 | Intelligent humanoid recognition alarm system and method based on monitoring camera |
| CN117853493A (en) * | 2024-03-08 | 2024-04-09 | 山东天意机械股份有限公司 | Wallboard production quality detection method based on image processing |
| CN117893527A (en) * | 2024-03-07 | 2024-04-16 | 江苏亿都智能特种装备有限公司 | New energy storage box surface defect detection method |
-
2024
- 2024-05-10 CN CN202410572336.9A patent/CN118154458B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020198997A1 (en) * | 2019-03-29 | 2020-10-08 | 深圳市大疆创新科技有限公司 | Connected domain processing method, data processing device and computer readable storage medium |
| CN114998112A (en) * | 2022-04-22 | 2022-09-02 | 南通悦福软件有限公司 | Image denoising method and system based on adaptive frequency domain filtering |
| CN116703910A (en) * | 2023-08-07 | 2023-09-05 | 威海丰荟建筑工业科技有限公司 | Intelligent detection method for quality of concrete prefabricated bottom plate |
| CN117237245A (en) * | 2023-11-16 | 2023-12-15 | 湖南云箭智能科技有限公司 | Industrial material quality monitoring method based on artificial intelligence and Internet of things |
| CN117809379A (en) * | 2024-02-29 | 2024-04-02 | 深圳市翔飞科技股份有限公司 | Intelligent humanoid recognition alarm system and method based on monitoring camera |
| CN117893527A (en) * | 2024-03-07 | 2024-04-16 | 江苏亿都智能特种装备有限公司 | New energy storage box surface defect detection method |
| CN117853493A (en) * | 2024-03-08 | 2024-04-09 | 山东天意机械股份有限公司 | Wallboard production quality detection method based on image processing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118154458B (en) | 2024-08-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116152239B (en) | Expressway pavement disease detection method and system | |
| CN114782475B (en) | Corrugated carton line pressing defect detection optimization method based on artificial intelligence system | |
| CN108711158B (en) | Pointer instrument image identification method based on contour fitting and radial segmentation | |
| CN115018838B (en) | Method for identifying pitting defects on surface of oxidized steel pipe material | |
| CN115471503B (en) | Equipment abnormity detection method for numerical control ingot splitting machine | |
| CN106780347B (en) | Early loquat bruise identification method based on OCT image processing | |
| CN116580024B (en) | Coke quality detection method based on image processing | |
| CN114139618A (en) | Signal dependent noise parameter estimation method based on improved density peak clustering | |
| CN114820625B (en) | Automobile top block defect detection method | |
| CN116309599B (en) | Water quality visual monitoring method based on sewage pretreatment | |
| CN117893527B (en) | New energy storage box surface defect detection method | |
| CN103942526B (en) | Linear feature extraction method for discrete data point set | |
| CN107145888A (en) | Video caption real time translating method | |
| CN114985283A (en) | Tea leaf identification method based on multi-feature image identification technology | |
| CN118154458B (en) | Filtering processing method of machine tool cutter image based on complex industrial environment | |
| CN117291843B (en) | Efficient management method for image database | |
| CN113723314A (en) | Sugarcane stem node identification method based on YOLOv3 algorithm | |
| CN115640546A (en) | Lithology identification method based on fusion of image and feature information | |
| CN113160166B (en) | Medical image data mining working method through convolutional neural network model | |
| CN113034403B (en) | Working method for denoising medical image data through Gaussian filtering | |
| CN114926360A (en) | Image noise reduction processing working method based on noise estimation | |
| CN118154456B (en) | Image intelligent denoising method for titanium alloy processing cutter model construction | |
| KR101408345B1 (en) | Data processing apparatus and method for classifying data, media for writing the method | |
| CN118247338B (en) | Accurate positioning method for numerical control machine tool machining | |
| CN113160167B (en) | Medical image data extraction working method through deep learning network model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |