CN114897890A - Artificial intelligence-based modified plastic production regulation and control method - Google Patents
Artificial intelligence-based modified plastic production regulation and control method Download PDFInfo
- Publication number
- CN114897890A CN114897890A CN202210797494.5A CN202210797494A CN114897890A CN 114897890 A CN114897890 A CN 114897890A CN 202210797494 A CN202210797494 A CN 202210797494A CN 114897890 A CN114897890 A CN 114897890A
- Authority
- CN
- China
- Prior art keywords
- defect
- connected domains
- possible defect
- domains
- same group
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Image Processing (AREA)
Abstract
The invention relates to the field of artificial intelligence, in particular to a modified plastic production regulation and control method based on artificial intelligence. The method comprises the following steps: acquiring a gray level image of the surface of the modified plastic, and acquiring edge pixel points according to edge probability; clustering edge pixel points to obtain a plurality of connected domains, acquiring every two possible defect connected domains with the distance value smaller than a second threshold value, and dividing every two corresponding possible defect connected domains into a group of connected domains; calculating the gradient direction similarity of each connected domain in the same group of connected domains and the direction consistency of the frameworks of every two possible defect connected domains; calculating the defect probability of each group of possible defect connected domains; calculating the continuous grain rate of the gray level image; and regulating and controlling the plastic production according to the grain connection rate. According to the method, the gray gradient similarity of each possible defect connected domain in the defect group and the consistency of every two possible defect connected domain frameworks in the group are calculated to judge, so that the judgment result is more accurate, and the precision control of the modified plastic production process is achieved.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a modified plastic production regulation and control method based on artificial intelligence.
Background
With the rapid increase of the consumption level of residents and the improvement of technologies of replacing steel with plastics and replacing copper with plastics, the demand of modified plastics is stably and rapidly increased, and the modified plastics belong to typical beneficial products of technical progress and consumption upgrade. The modified plastic product belongs to an intermediate product of a petrochemical industry chain, mainly takes five general plastics (acrylonitrile-butadiene-styrene ABS, polystyrene PS, polypropylene PP, polyethylene PE and polyvinyl chloride PVC) and five engineering plastics (polyamide PA, polycarbonate PC, polyformaldehyde POM, polybutylene terephthalate PBT and polyphenylene oxide PPO) as plastic base materials, is processed by adding fillers or additives with certain characteristics, and is widely applied to various fields of household appliances, automobiles, electric tools, office supplies, aerospace, war industry and the like at present.
The main production equipment of the modified plastic is a double-screw extruder, and the common production process flow comprises the steps of material preparation → material mixing → extrusion → bracing → cooling → grain cutting → packaging. The quality problems of modified plastic products can be mainly divided into appearance defects and unqualified performance, the appearance defects are mostly related to factors such as production process, raw material quality, production equipment, production environment and the like, and the common defects comprise long strips, continuous particles, black spots, discoloration, poor vacuum, scrap iron, poor plasticization, carbonization and the like.
Wherein the grain connection defect is the most common, two or more grains are connected in parallel to be called as grain connection, the reason for the grain connection defect is usually caused by strip breakage during strip pulling or insufficient water passing length, the modified plastic in a molten state is subjected to strip pulling through a double-screw extruder, the strip pulling is extruded to enter a strip guide wheel, a plurality of extruded material strips are separated, when the width of the selected guide bar wheel is not proper, the adjacent material bars are adhered, two or more particles are adhered together in parallel to form a continuous particle, meanwhile, the length of improper cooling water passing can also cause particle connection, too short length of cooling water passing can cause too soft extrusion bracing strips, adjacent material strips are adhered together under the pressure action of the granulator when the granulator cuts, and too long length of cooling water passing can cause too large rigidity of materials, the broken strip is caused under the action of unmatched traction force, so that the reasonable width of the guide wheel strip and the cooling water passing length are important.
Aiming at the situation, the invention provides a modified plastic production regulation and control method based on artificial intelligence by analyzing grain cutting images collected on a conveyor belt, obtains grain connection rate by identifying the condition of grain connection particles in the images, and controls the width of a guide wheel strip and the length of cooling water according to the grain connection rate.
Disclosure of Invention
The invention provides a modified plastic production regulation and control method based on artificial intelligence, which aims to solve the existing problems and comprises the following steps:
collecting a gray level image of the surface of the modified plastic, and marking pixel points with edge probability greater than a first threshold value as edge pixel points; selecting edge pixel points to perform clustering to obtain a plurality of connected domains, acquiring pairwise possible defect connected domains with the distance values smaller than a second threshold value, and dividing the pairwise possible defect connected domains into a group; calculating the gradient direction similarity of each connected domain in the same group and the direction consistency of the skeletons of every two possible defect connected domains; calculating the defect probability of each group of possible defect connected domains; calculating the continuous grain rate of the gray level image; and regulating and controlling the plastic production according to the grain connection rate.
According to the technical means provided by the invention, the traditional distance field-based framework extraction algorithm is improved, so that the extracted framework is more accurate, the calculated grain connection rate is more precise and accurate, the judgment result is more accurate by calculating the gray gradient similarity of each possible defect connected domain in the defect group and the consistency of every two possible defect connected domain frameworks in the group, and the accurate control of the modified plastic production process is achieved.
The invention adopts the following technical scheme that a modified plastic production regulation and control method based on artificial intelligence comprises the following steps:
and acquiring a modified plastic surface gray image, and acquiring the gradient amplitude and the gradient direction of each pixel point in the modified plastic surface gray image.
And acquiring all edge pixel points in the modified plastic surface gray level image according to the gradient amplitude of each pixel point.
And clustering according to the edge pixel points to obtain a plurality of connected domains, screening out possible defect connected domains according to the distance between the central point coordinates of every two connected domains, and dividing every two possible defect connected domains with the distance value smaller than a second threshold value into a group of connected domains.
And calculating the gradient direction similarity of each connected domain in the group according to the gradient directions of all the pixel points in each possible defect connected domain in the same group of connected domains.
And extracting the skeleton of each possible defect connected domain in the same group of connected domains, and calculating the direction consistency of every two possible defect connected domain skeletons in the same group of connected domains.
And calculating the defect probability of the same group of connected domains belonging to the defect connected domains according to the gradient direction similarity of all possible defect connected domains in the same group of connected domains, the direction consistency of skeletons of every two possible defect connected domains in the same group of connected domains and the distance of coordinates of central points of every two possible defect connected domains in the same group of connected domains.
And judging whether the group is a defect connected domain group according to the defect probability of the same group belonging to the defect connected domain, and regulating and controlling the plastic production according to the determined number of the defect connected domain groups.
Further, a method for regulating and controlling production of modified plastics based on artificial intelligence, which comprises the following steps of clustering according to the edge pixel points:
selecting the edge pixel point with the maximum edge probability as a growth seed point by using a region growth method, searching other edge pixel points in eight neighborhoods of the growth seed point, and combining other pixel points in the eight neighborhoods with the growth seed point;
repeating iteration by taking the edge pixel points in the eight neighborhoods as new growth seed points until no edge pixel points exist in the eight neighborhoods of the new growth seed points, and obtaining a clustering result;
and repeating the steps for other edge pixel points to obtain a plurality of clustering results, wherein each clustering result corresponds to a connected domain.
Further, a method for producing and regulating modified plastics based on artificial intelligence, which comprises the following steps of:
acquiring the direction from each internal point to the boundary point in each possible defect connected domain, and calculating the gradient amplitude mean value from each internal point to the corresponding connected domain boundary point in each possible defect connected domain; the boundary point is the closest point to the corresponding interior point, and the direction from the interior point to the boundary point is the direction average value of each interior point and the boundary point;
and screening all internal points of which the gradient amplitude is larger than a preset threshold value in each possible defect connected domain to obtain a skeleton of each possible defect connected domain.
Further, a method for regulating and controlling the production of modified plastics based on artificial intelligence, which is used for calculating the direction consistency of the frameworks of every two possible defect connected domains in the same group of connected domains, and comprises the following steps:
fitting the skeleton of each connected domain in the same group by using a least square method to obtain a direction vector of each connected domain skeleton, and calculating the expression of the direction consistency as follows:
in the formula
The direction consistency is represented by the direction consistency,
represents the direction vector of the skeleton of connected component 1,
representing the direction vector of the skeleton of connected component 2.
Further, an artificial intelligence based modified plastic production regulation and control method, wherein the expression for calculating the defect probability of the possible defect connected domain in the group is as follows:
wherein the content of the first and second substances,
the probability of the defect is represented by,
distance values representing coordinates of center points of two possible defect connected domains in the same group,
representing the gradient direction similarity of the ith possible defect connected domain in the same group, n representing n possible defect connected domains in the same group,
representing the direction consistency of the skeletons of every two possible defect connected domains in the same group.
Further, the method for regulating and controlling the production of the modified plastic based on the artificial intelligence counts the number of possible defect connected domains in a group with the probability of all defects being larger than a third threshold value in the gray level image of the surface of the modified plastic, and obtains the grain connection rate of the gray level image of the surface of the modified plastic according to the ratio of the number of all defect connected domains in the group with the probability of the defects being larger than the third threshold value to the number of all connected domains in the gray level image of the surface of the modified plastic.
Further, the method for regulating and controlling the production of the modified plastics based on the artificial intelligence comprises the following steps of:
wherein the content of the first and second substances,
the width of the guide wheel strip at the time of acquisition is shown,
indicating the length of cooling water at the time of acquisition,
indicating the width of the guide wheel strip at the current time after adjustment,
showing the cooling water length at the current moment after adjustment, k showing the grain connection rate of the plastic grain connection defect gray level image at the acquisition moment,
is a hyper-parameter.
The invention has the beneficial effects that: according to the technical means provided by the invention, the traditional distance field-based framework extraction algorithm is improved, so that the extracted framework is more accurate, the calculated grain connection rate is more precise and accurate, the judgment result is more accurate by calculating the gray gradient similarity of each possible defect connected domain in the defect group and the consistency of every two possible defect connected domain frameworks in the group, and the accurate control of the modified plastic production process is achieved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a schematic structural diagram of a modified plastic production control method based on artificial intelligence according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1, a schematic structural diagram of a modified plastic production regulation method based on artificial intelligence in an embodiment of the present invention is provided, which includes:
101. and acquiring a modified plastic surface gray image, and acquiring the gradient amplitude and the gradient direction of each pixel point in the modified plastic surface gray image.
The method needs to identify the existence and distribution condition of the continuous particles on the surface of the modified plastic particles in the conveyor belt area, and needs to collect the surface gray image of the conveyor belt area to be detected. Arranging a camera, wherein the camera is arranged at a position 300mm away from the upper end of the transmission roller platform, and the lens and the roller platform are arranged at an angle of 90 degrees.
The light source is a strip-shaped white LED light source, the LED light source and the conveyor belt are arranged at an angle of 45 degrees, the modified plastic particle images on the surface of the conveyor belt are collected once every fixed time, and the speed of the conveyor belt is
S denotes the length of the conveyor belt covered by the image acquisition and t denotes the sampling interval time.
In order to reduce unnecessary calculation caused by environmental factors outside a conveyor belt, the invention firstly adopts a DNN technology to identify the images to be detected of the conveyor belt in the images, and the relevant content of the DNN network is as follows:
the data set used is the data set of the surface image of the area of the conveyor belt to be detected obtained in the acquisition process of the invention, and the distribution pattern and the form of the modified plastic particles on the conveyor belt to be detected are various.
The pixels needing to be segmented are divided into two types, namely the labeling process of the corresponding labels of the training set is as follows: and in the single-channel semantic label, the label of the pixel at the corresponding position belonging to the background class is 0, and the label of the pixel belonging to the surface image of the to-be-detected conveyor belt area is 1.
The task of the network is to classify, and all the used loss functions are cross entropy loss functions.
Obtaining modified plastic particle edge pixel points, wherein the modified plastic particle boundary is obvious because the modified plastic particle is brighter and the conveyor belt is darker, and calculating all pixel points in the image by using a sobel operator
，
Gradient of direction
Then the gradient amplitude of the pixel point
Obtaining the gradient amplitude and gradient direction of all pixel points of the whole image, wherein the corresponding gradient direction is
。
102. And acquiring all edge pixel points in the modified plastic surface gray level image according to the gradient amplitude of each pixel point.
Because the modified plastic particle is a cylinder, gray gradient exists in the modified plastic particle, and the gray gradient of the modified plastic edge pixel point is larger than the gray gradient of the modified plastic edge pixel point, so that:
wherein Q represents the probability of the edge pixel, and c represents the gray gradient value most likely to be the edge, wherein
Respectively representing the minimum and maximum values of the gray gradient,
gradient amplitude representing pixel point, C =
（
-
) And when the value of Q is larger than 0.8, marking the pixel point as an edge pixel point.
103. And clustering according to the edge pixel points to obtain a plurality of connected domains, screening out possible defect connected domains according to the distance between the central point coordinates of every two connected domains, and dividing every two possible defect connected domains with the distance value smaller than a second threshold value into a group of connected domains.
The invention adopts a region growing method based on edge probability to perform clustering, selects the pixel point with the maximum edge probability from the marked pixel points as a growing seed point, randomly selects one of the marked pixel points if the marked pixel points have the maximum edge probability values, searches in 8 neighborhoods of the selected pixel point, and the pixel points in the neighborhoods belong to the edge probability
The pixel points of (1) are reserved and combined into a region.
The region is used as a new growth seed point, searching is carried out in the neighborhood again, and the pixel points in the neighborhood belong to the marginal probability
The pixel points are reserved, the area is updated to obtain a new seed point area, and iteration is carried out for multiple times until the neighborhood does not contain the edge probability
The pixel point of (2) is stopped, and the first connected domain is obtained at the moment.
For the rest pixel points, selecting the pixel point with the maximum edge probability as a growth seed point in the same way, if the maximum edge probability exists in the plurality of edge probabilities, randomly selecting one of the pixel points, searching in 8 neighborhoods of the selected pixel point, wherein the pixel point in the neighborhood belongs to the edge probability
The pixel points of (1) are reserved and combined into a region.
At the moment, the area is taken as a new growth seedAnd (4) searching the sub-points in the neighborhood again, wherein the edge probability of the pixel points in the neighborhood
The pixel points are reserved, a new region is updated to obtain a new seed point region, and iteration is carried out for multiple times until the neighborhood does not contain the edge probability
And stopping the process of obtaining the second connected domain.
Repeating the operation on the rest pixel points, and iterating to obtain a plurality of connected domains until the edge probability
The clustering is completed, and a plurality of connected domains are obtained at the moment.
The method for clustering by taking any edge pixel point as a center to obtain a plurality of connected domains comprises the following steps:
selecting the edge pixel point with the maximum edge probability as a growth seed point by using a region growth method, searching other edge pixel points in eight neighborhoods of the growth seed point, and combining other pixel points in the eight neighborhoods with the growth seed point;
repeating iteration by taking the edge pixel points in the eight neighborhoods as new growth seed points until no edge pixel points exist in the eight neighborhoods of the new growth seed points, and obtaining a clustering result;
and repeating the steps for other edge pixel points to obtain a plurality of clustering results, wherein each clustering result corresponds to a connected domain.
And when every two corresponding possible defect connected domains are divided into a group, if the distance value of the central point coordinates of a plurality of connected domains is smaller than a second threshold value, the corresponding plurality of connected domains are divided into a group.
Acquiring the central points of the connected domains, and calculating the distance between every two central points of the connected domains:
wherein L represents the distance between two connected center points of the domain, (L)
，
），（
，
) And when L is less than or equal to 20 (pixel is unit), the two connected domains are connected domain with particle possible defects, the connected domains are marked, and the connected domains meeting the threshold requirement are divided into a group.
104. And calculating the similarity of the gradient direction of each connected domain in the same group of connected domains according to the gradient direction of all the pixel points in each possible defect connected domain in the same group of connected domains.
For overlapped modified plastic particles, the gradient direction consistency of pixel points in a connected domain is poor due to the change of the pose, and the gradient direction consistency of the pixel points in the connected domain in normal or connected particles is good due to the structure of a cylinder, so that:
in the formula
The similarity of the gradient directions of the pixel points in the connected domain is expressed, Z represents the number of the pixel points in the connected domain,
the gradient direction of the z-th pixel point in the connected domain is represented,
and expressing the gradient direction of the s-th pixel point in four neighborhoods of the z-th pixel point in the connected domain, wherein s expresses the number of the neighborhood pixel points of the z-th pixel point.
105. And extracting the skeleton of each possible defect connected domain in the same group of connected domains, and calculating the direction consistency of every two possible defect connected domain skeletons in the same group of connected domains.
The traditional framework extraction algorithm based on the distance field can extract the central framework of the connected domain, but the central framework is not necessarily an ideal real framework due to the change of the pose of the modified plastic particles, so the particle framework is extracted by adopting a gradient amplitude framework extraction method.
The method for extracting the skeleton of each possible defect connected domain in the same group comprises the following steps:
acquiring the direction from each internal point to the boundary point in each possible defect connected domain, and calculating the gradient amplitude mean value from each internal point to the corresponding connected domain boundary point in each possible defect connected domain; the boundary point is the closest point to the corresponding interior point, and the direction from the interior point to the boundary point is the direction average value of each interior point and the boundary point;
and screening all internal points of which the gradient amplitude is larger than a preset threshold value in each possible defect connected domain to obtain a skeleton of each possible defect connected domain.
The method for calculating the direction consistency of every two possible defect connected domain frameworks in the same group comprises the following steps:
fitting the skeleton of each connected domain in the same group by using a least square method to obtain a direction vector of each connected domain skeleton, and calculating the expression of the direction consistency as follows:
in the formula
The direction consistency is represented by the direction consistency,
represents the direction vector of the skeleton of connected component 1,
representing the direction vector of the skeleton of connected component 2. The more consistent the main directions of the frameworks of every two connected domains are, the more likely the two corresponding connected domains are the grain-connected defect.
106. And calculating the defect probability of the same group of connected domains belonging to the defect connected domains according to the gradient direction similarity of all possible defect connected domains in the same group of connected domains, the direction consistency of skeletons of every two possible defect connected domains in the same group of connected domains and the distance of coordinates of central points of every two possible defect connected domains in the same group of connected domains.
The expression for calculating the defect probability of each group of possible defect connected domains is as follows:
wherein the content of the first and second substances,
the probability of the defect is represented by,
distance values representing coordinates of center points of two possible defect connected domains in the same group,
representing the gradient direction similarity of the ith possible defect connected domain in the same group, n representing n possible defect connected domains in the same group,
representing the direction consistency of the skeletons of every two possible defect connected domains in the same group. In the invention
When the connected component is a possible defect connected component, the connected component in the group is a possible defect connected component.
107. And judging whether the group is a defect connected domain group according to the defect probability of the same group belonging to the defect connected domain, and regulating and controlling the plastic production according to the determined number of the defect connected domain groups.
And counting the number of defect connected domains in the group with the defect probability larger than a third threshold value in the modified plastic surface gray level image, and obtaining the grain connection rate of the modified plastic surface gray level image according to the ratio of the number of all defect connected domains in the group with the defect probability larger than the third threshold value to the number of all connected domains in the modified plastic surface gray level image.
Counting the number of all connected domains in the image, wherein the number is the number of the modified plastic particles collected in the image and is recorded as
Counting the number of connected domains of particle-linked defects, and recording the number as
And then the grain-linking rate k of the grain-linking defects in the image is:
the larger the value of the grain rate k is, the larger the grain rate k is, the grain is connected due to the fact that the width of the guide strip wheel is too small to cause adhesion of adjacent brace bars during the strip extrusion, the length of the material strip cooling water is insufficient, the material strip cannot be sufficiently cooled, the phenomenon that the strip extrusion is too soft to cause grain connection is caused, and therefore adjusting force is large.
The width of the guide wheel is unreasonable when the grain rate of the modified plastic is higher than that of the extrusion bracing strip, the material strips connected together cannot be separated, the length of the material strip cooling water is insufficient, the material strips cannot be sufficiently cooled, the extrusion bracing strip is too soft, the adjacent material strips are adhered together under the pressure action of the granulator when the granulator cuts, and the width of the guide wheel strip and the length of the cooling water need to be regulated and controlled at the moment, so that the grain rate of the guide wheel strip is reduced.
The higher the continuous grain rate is, the wider the guide wheel strip width and the longer the cooling water length are needed on the current basis, and then the method for regulating and controlling the plastic production according to the continuous grain rate is as follows:
wherein the content of the first and second substances,
the width of the guide wheel strip at the time of acquisition is shown,
the length of the cooling water at the time of collection is shown,
indicating the width of the guide wheel strip at the current time after adjustment,
showing the cooling water length at the current moment after adjustment, k showing the grain connection rate of the plastic grain connection defect gray level image at the acquisition moment,
is a hyper-parameter.
The width of the guide wheel strip and the length of the cooling water are regulated and controlled according to the grain connection rate at the collection time, so that the grain connection rate at the current time after regulation is reduced, and the quality and the qualified rate of the modified plastic particles are improved.
According to the technical means provided by the invention, the traditional distance field-based framework extraction algorithm is improved, so that the extracted framework is more accurate, the calculated grain connection rate is more precise and accurate, the judgment result is more accurate by calculating the gray gradient similarity of each possible defect connected domain in the defect group and the consistency of every two possible defect connected domain frameworks in the group, and the accurate control of the modified plastic production process is achieved.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (6)
1. A production regulation and control method of modified plastics based on artificial intelligence is characterized by comprising the following steps:
acquiring a modified plastic surface gray image, and acquiring the gradient amplitude and the gradient direction of each pixel point in the modified plastic surface gray image;
acquiring all edge pixel points in the modified plastic surface gray level image according to the gradient amplitude of each pixel point;
clustering is carried out according to the edge pixel points to obtain a plurality of connected domains, possible defect connected domains are screened out according to the distance between the coordinates of the central points of every two connected domains, and every two possible defect connected domains with the distance value smaller than a second threshold value are divided into a group of connected domains;
calculating the gradient direction similarity of each connected domain in the group according to the gradient directions of all pixel points in each possible defect connected domain in the same group of connected domains;
extracting the skeleton of each possible defect connected domain in the same group of connected domains, and calculating the direction consistency of every two possible defect connected domain skeletons in the same group of connected domains;
calculating the defect probability of the same group of connected domains belonging to the defect connected domains according to the gradient direction similarity of all possible defect connected domains in the same group of connected domains, the direction consistency of frameworks of every two possible defect connected domains in the same group of connected domains and the distance of coordinates of central points of every two possible defect connected domains in the same group of connected domains;
the expression for calculating the defect probability of the possible defect connected domain in the same group is as follows:
wherein the content of the first and second substances,
to representThe probability of the defect is such that,
distance values representing coordinates of center points of two possible defect connected domains in the same group,
representing the gradient direction similarity of the ith possible defect connected domain in the same group, n representing n possible defect connected domains in the same group,
representing the direction consistency of every two possible defect connected domain skeletons in the same group;
and judging whether the group is a defect connected domain group according to the defect probability of the same group belonging to the defect connected domain, and regulating and controlling the plastic production according to the determined number of the defect connected domain groups.
2. The artificial intelligence-based modified plastic production regulation and control method as claimed in claim 1, wherein the method for clustering according to the edge pixel points to obtain a plurality of connected domains comprises:
selecting the edge pixel point with the maximum edge probability as a growth seed point by using a region growth method, searching other edge pixel points in eight neighborhoods of the growth seed point, and combining other pixel points in the eight neighborhoods with the growth seed point;
repeating iteration by taking the edge pixel points in the eight neighborhoods as new growth seed points until no edge pixel points exist in the eight neighborhoods of the new growth seed points, and obtaining a clustering result;
and repeating the steps for other edge pixel points to obtain a plurality of clustering results, wherein each clustering result corresponds to a connected domain.
3. The artificial intelligence-based modified plastic production regulation and control method as claimed in claim 1, wherein the method for extracting the skeleton of each possible defect connected domain in the same group comprises:
acquiring the direction from each internal point to the boundary point in each possible defect connected domain, and calculating the gradient amplitude mean value from each internal point to the corresponding connected domain boundary point in each possible defect connected domain; the boundary point is the closest point to the corresponding internal point, and the direction from the internal point to the boundary point is the direction average value of each internal point and the boundary point;
and screening all internal points of which the gradient amplitude is larger than a preset threshold value in each possible defect connected domain to obtain a skeleton of each possible defect connected domain.
4. The artificial intelligence-based modified plastic production regulation and control method as claimed in claim 3, wherein the method for calculating the direction consistency of the frameworks of every two possible defective connected domains in the same group of connected domains comprises:
fitting the skeleton of each connected domain in the same group by using a least square method to obtain a direction vector of each connected domain skeleton, and calculating the expression of the direction consistency as follows:
in the formula
The direction consistency is represented by the direction consistency,
represents the direction vector of the skeleton of connected component 1,
representing the direction vector of the skeleton of connected component 2.
5. The artificial intelligence-based modified plastic production regulation and control method as claimed in claim 1, wherein the number of possible defect connected domains in the group with the defect probability greater than the third threshold value in the modified plastic surface gray level image is counted, and the grain connection rate of the modified plastic surface gray level image is obtained by the ratio of the number of all defect connected domains in the group with the defect probability greater than the third threshold value to the number of all connected domains in the modified plastic surface gray level image.
6. The artificial intelligence-based modified plastic production regulating method as claimed in claim 5, wherein the method for regulating the plastic production according to the determined number of defect connected domain groups comprises:
wherein the content of the first and second substances,
the width of the guide wheel strip at the time of acquisition is shown,
the length of the cooling water at the time of collection is shown,
indicating the width of the guide wheel strip at the current time after adjustment,
showing the cooling water length at the current moment after adjustment, k showing the grain connection rate of the plastic grain connection defect gray level image at the acquisition moment,
is a hyper-parameter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210797494.5A CN114897890B (en) | 2022-07-08 | 2022-07-08 | Artificial intelligence-based modified plastic production regulation and control method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210797494.5A CN114897890B (en) | 2022-07-08 | 2022-07-08 | Artificial intelligence-based modified plastic production regulation and control method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114897890A true CN114897890A (en) | 2022-08-12 |
| CN114897890B CN114897890B (en) | 2022-09-30 |
Family
ID=82729865
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210797494.5A Active CN114897890B (en) | 2022-07-08 | 2022-07-08 | Artificial intelligence-based modified plastic production regulation and control method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114897890B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114998321A (en) * | 2022-07-19 | 2022-09-02 | 南通博纳纺织品有限公司 | Textile material surface hairiness degree identification method based on optical means |
| CN116030060A (en) * | 2023-03-29 | 2023-04-28 | 高唐县红发塑业有限公司 | Plastic particle quality detection method |
| CN116152255A (en) * | 2023-04-21 | 2023-05-23 | 高唐县红发塑业有限公司 | Modified plastic production defect judging method |
| CN116630332A (en) * | 2023-07-26 | 2023-08-22 | 山东华航高分子材料有限公司 | PVC plastic pipe orifice defect detection method based on image processing |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130202188A1 (en) * | 2012-02-06 | 2013-08-08 | Hitachi High-Technologies Corporation | Defect inspection method, defect inspection apparatus, program product and output unit |
| CN113888536A (en) * | 2021-12-01 | 2022-01-04 | 武汉市利友印务有限公司 | Printed matter double image detection method and system based on computer vision |
| CN114419025A (en) * | 2022-01-27 | 2022-04-29 | 江苏泰和木业有限公司 | Fiberboard quality evaluation method based on image processing |
| CN114445387A (en) * | 2022-01-29 | 2022-05-06 | 泗阳富艺木业股份有限公司 | Fiberboard quality classification method based on machine vision |
| CN114670421A (en) * | 2022-05-31 | 2022-06-28 | 南通净缘塑料制品有限公司 | Machine vision-based modified plastic extrusion production control method |
-
2022
- 2022-07-08 CN CN202210797494.5A patent/CN114897890B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130202188A1 (en) * | 2012-02-06 | 2013-08-08 | Hitachi High-Technologies Corporation | Defect inspection method, defect inspection apparatus, program product and output unit |
| CN113888536A (en) * | 2021-12-01 | 2022-01-04 | 武汉市利友印务有限公司 | Printed matter double image detection method and system based on computer vision |
| CN114419025A (en) * | 2022-01-27 | 2022-04-29 | 江苏泰和木业有限公司 | Fiberboard quality evaluation method based on image processing |
| CN114445387A (en) * | 2022-01-29 | 2022-05-06 | 泗阳富艺木业股份有限公司 | Fiberboard quality classification method based on machine vision |
| CN114670421A (en) * | 2022-05-31 | 2022-06-28 | 南通净缘塑料制品有限公司 | Machine vision-based modified plastic extrusion production control method |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114998321A (en) * | 2022-07-19 | 2022-09-02 | 南通博纳纺织品有限公司 | Textile material surface hairiness degree identification method based on optical means |
| CN116030060A (en) * | 2023-03-29 | 2023-04-28 | 高唐县红发塑业有限公司 | Plastic particle quality detection method |
| CN116152255A (en) * | 2023-04-21 | 2023-05-23 | 高唐县红发塑业有限公司 | Modified plastic production defect judging method |
| CN116630332A (en) * | 2023-07-26 | 2023-08-22 | 山东华航高分子材料有限公司 | PVC plastic pipe orifice defect detection method based on image processing |
| CN116630332B (en) * | 2023-07-26 | 2023-09-26 | 山东华航高分子材料有限公司 | PVC plastic pipe orifice defect detection method based on image processing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114897890B (en) | 2022-09-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114897890B (en) | Artificial intelligence-based modified plastic production regulation and control method | |
| CN104624505B (en) | A kind of waste plastics separation method based on image recognition and system | |
| CN114445387A (en) | Fiberboard quality classification method based on machine vision | |
| CN102184405B (en) | Image acquisition-analysis method | |
| CN108181316B (en) | Bamboo strip defect detection method based on machine vision | |
| CN104574389A (en) | Battery piece chromatism selection control method based on color machine vision | |
| CN110667148B (en) | Film thickness control system | |
| CN107330365A (en) | Traffic sign recognition method based on maximum stable extremal region and SVM | |
| CN102609723A (en) | Image classification based method and device for automatically segmenting videos | |
| CN109063667A (en) | A kind of video identification method optimizing and method for pushing based on scene | |
| CN114972894A (en) | CPP film defect classification method based on computer vision | |
| CN111932639A (en) | Detection method of unbalanced defect sample based on convolutional neural network | |
| Rahman et al. | Identification of mature grape bunches using image processing and computational intelligence methods | |
| CN112446417B (en) | Spindle-shaped fruit image segmentation method and system based on multilayer superpixel segmentation | |
| CN204148153U (en) | A kind of multichannel color selector | |
| Huang et al. | Surface defects detection for mobilephone panel workpieces based on machine vision and machine learning | |
| CN110728677B (en) | Texture roughness defining method based on sliding window algorithm | |
| CN107310133A (en) | BOPP film method for controlling thickness based on image procossing | |
| CN114953306B (en) | Color composite rubber belt calendering process quality control system | |
| CN111709919A (en) | Glass defect visual detection algorithm | |
| El-Labban et al. | Discriminative semi-markov models for automated mitotic phase labelling | |
| CN113560214B (en) | Intelligent sorting system for pearls and application method thereof | |
| CN113705531A (en) | Method for identifying alloy powder inclusions based on microscopic imaging | |
| CN117094609B (en) | Intelligent management system for aluminum profile production quality based on machine vision | |
| CN110766673B (en) | Texture roughness defining method based on Euclidean distance judgment |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230224 Address after: No. 9, Yuelong Road, Qidong Science and Technology Innovation Park, Nantong, Jiangsu 226200 Patentee after: JIANGSU QIHANG LUGGAGE Co.,Ltd. Address before: 226200 No. 18, cailong Middle Road, Qidong science and technology entrepreneurship Park, Nantong City, Jiangsu Province Patentee before: Nantong Huaye Plastic Industry Co.,Ltd. |