CN118961756B - A method for detecting sealing failure of biological products based on image recognition - Google Patents
A method for detecting sealing failure of biological products based on image recognition Download PDFInfo
- Publication number
- CN118961756B CN118961756B CN202411462732.2A CN202411462732A CN118961756B CN 118961756 B CN118961756 B CN 118961756B CN 202411462732 A CN202411462732 A CN 202411462732A CN 118961756 B CN118961756 B CN 118961756B
- Authority
- CN
- China
- Prior art keywords
- defect
- area
- sealing
- false
- biological product
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000007789 sealing Methods 0.000 title claims abstract description 125
- 238000000034 method Methods 0.000 title claims abstract description 39
- 230000007547 defect Effects 0.000 claims abstract description 247
- 238000001514 detection method Methods 0.000 claims abstract description 38
- 238000004458 analytical method Methods 0.000 claims abstract description 35
- 239000003566 sealing material Substances 0.000 claims abstract description 35
- 238000012545 processing Methods 0.000 claims abstract description 11
- 230000003213 activating effect Effects 0.000 claims abstract description 4
- 238000009826 distribution Methods 0.000 claims description 34
- 238000007621 cluster analysis Methods 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 6
- 238000005516 engineering process Methods 0.000 abstract description 4
- 238000013527 convolutional neural network Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 2
- 230000002950 deficient Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000037303 wrinkles Effects 0.000 description 2
- 230000015556 catabolic process Effects 0.000 description 1
- 238000011109 contamination Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000000813 microbial effect Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 229960005486 vaccine Drugs 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/95—Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
本发明公开了基于图像识别的生物制品密封失效检测方法，涉及密封检测技术领域。所述方法包括：通过第一柔性密封材料和第二柔性密封材料对生物制品进行密封并传输至第一检测平台，生物制品在所述第一检测平台具有密封区域定位坐标；激活摄像头，基于密封区域定位坐标，采集生物制品的密封区域图像信息；通过缺陷识别器，对密封区域图像信息进行处理，提取假性缺陷区域；对假性缺陷区域进行纹理一致性分析，获得选定缺陷区域；当选定缺陷区域的尺寸规格满足失效尺寸规格，对生物制品进行密封失效标识。解决了现有技术中生物制品密封失效检测准确性较差的技术问题，通过图像识别技术，达到了提高检测效率和准确性的技术效果。
The present invention discloses a method for detecting sealing failure of biological products based on image recognition, and relates to the technical field of sealing detection. The method comprises: sealing the biological product by a first flexible sealing material and a second flexible sealing material and transmitting the sealing product to a first detection platform, and the biological product has a sealing area positioning coordinate on the first detection platform; activating a camera, and collecting the sealing area image information of the biological product based on the sealing area positioning coordinate; processing the sealing area image information by a defect identifier, and extracting the false defect area; performing texture consistency analysis on the false defect area to obtain the selected defect area; when the size specification of the selected defect area meets the failure size specification, marking the sealing failure of the biological product. The method solves the technical problem of poor accuracy of sealing failure detection of biological products in the prior art, and achieves the technical effect of improving detection efficiency and accuracy through image recognition technology.
Description
Technical Field
The invention relates to the technical field of seal detection, in particular to a biological product seal failure detection method based on image recognition.
Background
The integrity of the seal is critical to ensure the safety, effectiveness and stability of the product during the production and storage of the bioproduct. Biological products, including but not limited to pharmaceuticals, vaccines, biotechnology products, and the like, often require storage and transportation in a sterile and sealed environment to prevent microbial contamination and product degradation. However, defects in the sealing material or errors in the sealing process may cause the sealing to fail, thereby affecting the quality and safety of the product. Conventional seal inspection methods generally employ a defect recognition model to identify defects in non-seal junction areas, however, seal port areas are often accompanied by complex wrinkles and textures, which make it difficult for a generic defect recognition model to accurately distinguish between true defects and normal structural features, e.g., natural wrinkles of a seal port may be misidentified as defects, while true micro-cracks or leaks may be missed due to insensitivity of the model.
Disclosure of Invention
The application provides a biological product sealing failure detection method based on image recognition, which solves the technical problem of poor accuracy of biological product sealing failure detection in the prior art.
In view of the above, the present application provides a method for detecting seal failure of a biological product based on image recognition.
The application provides a biological product sealing failure detection method based on image recognition, which comprises the following steps:
The method comprises the steps of sealing a biological product through a first flexible sealing material and a second flexible sealing material and transmitting the biological product to a first detection platform, wherein the biological product is provided with a sealing area positioning coordinate on the first detection platform, a sealing area is a sealing junction area of the first flexible sealing material and the second flexible sealing material, a camera is activated, sealing area image information of the biological product is collected based on the sealing area positioning coordinate, a defect identifier is used for processing the sealing area image information, a false defect area is extracted, texture consistency analysis is carried out on the false defect area, a selected defect area is obtained, and when the size specification of the selected defect area meets a failure size specification, a sealing failure identification is carried out on the biological product.
One or more technical schemes provided by the application have at least the following technical effects or advantages:
Firstly, sealing a biological product through a first flexible sealing material and a second flexible sealing material and transmitting the biological product to a first detection platform, wherein the biological product is provided with a sealing area positioning coordinate on the first detection platform, and the sealing area is a sealing interface area of the first flexible sealing material and the second flexible sealing material. Then, the camera is activated, and sealing area image information of the biological product is acquired based on the sealing area positioning coordinates. Then, the seal region image information is processed by a defect identifier to extract a pseudo defect region. And finally, carrying out texture consistency analysis on the pseudo defect area to obtain a selected defect area, and carrying out seal failure identification on the biological product when the size specification of the selected defect area meets the failure size specification. The technical problem of biological product seal failure detection accuracy is relatively poor in the prior art is solved, and the technical effects of improving detection efficiency and accuracy are achieved through an image recognition technology.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following description will briefly explain the drawings needed in the description of the embodiments, which are merely examples of the present invention, and other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of a biological product seal failure detection method based on image recognition according to an embodiment of the present application;
fig. 2 is a schematic flow chart of texture consistency analysis in the method for detecting seal failure of biological product based on image recognition according to the embodiment of the application.
Detailed Description
The application solves the technical problem of poor accuracy of biological product seal failure detection in the prior art by providing the biological product seal failure detection method based on image recognition.
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. It will be apparent that the described embodiments are only some, but not all, embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
It should be noted that the terms "comprises" and "comprising" are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or server that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or apparatus.
An embodiment, as shown in fig. 1, provides a method for detecting seal failure of a biological product based on image recognition, wherein the method comprises the following steps:
And sealing the biological product through a first flexible sealing material and a second flexible sealing material and transmitting the biological product to a first detection platform, wherein the biological product is provided with a sealing area positioning coordinate on the first detection platform, and the sealing area is a sealing junction area of the first flexible sealing material and the second flexible sealing material.
In the sealing detection process of biological products, first, a first flexible sealing material and a second flexible sealing material are adopted to seal the biological products, and after sealing is finished, the biological products are transmitted to a special first detection platform. Each biologic is assigned a precise seal area location coordinate on the first test platform that indicates the location of the seal area where the first flexible seal material and the second flexible seal material interface.
Further, the biological product is sealed by a first flexible sealing material and a second flexible sealing material and is transmitted to a first detection platform, the biological product has a sealing area positioning coordinate on the first detection platform, and the method comprises the following steps:
The method comprises the steps of obtaining sealing control parameters of a first flexible sealing material and a second flexible sealing material on the biological product, wherein the sealing control parameters are provided with coordinate feature identifiers, carrying out region clustering analysis on the coordinate feature identifiers according to the sealing control parameters to obtain region clustering results, wherein the same category of the region clustering results is provided with the same sealing control parameters, and adding the region clustering results into the sealing region positioning coordinates.
The method comprises the steps of obtaining key control parameters when a first flexible sealing material and a second flexible sealing material seal biological products, wherein the key control parameters not only reflect physical and chemical conditions of a sealing process, but also have coordinate characteristic identifiers, namely, the corresponding relation between the sealing control parameters and a sealing area in space, performing area clustering analysis by utilizing the sealing control parameters with the coordinate characteristic identifiers so as to divide data points with similar characteristics into the same category, optionally, clustering different coordinate characteristic identifiers by adopting K-means clustering according to the similarity of the sealing control parameters to form area clustering results, wherein the area clustering results of the same category have the same sealing control parameters, and integrating the area clustering results into positioning coordinates of the sealing area. By combining the region clustering result with the seal region positioning coordinates, the seal effect can be evaluated more accurately.
And activating a camera, and acquiring sealing area image information of the biological product based on the positioning coordinates of the sealing area.
Activating the high-precision camera, and acquiring image information of the sealing area of the biological product by utilizing the positioning coordinates of the sealing area. The image acquisition is carried out based on the coordinate information, so that the accuracy and pertinence of the image data are ensured, and a solid foundation is laid for subsequent analysis and processing.
And processing the image information of the sealing area through a defect identifier to extract a false defect area.
In the process of biological product sealing detection, the sealing area image information acquired by the camera is sent to a defect identifier for deep processing, so that an area with obvious difference from a normal area, namely a false defect area, is extracted, wherein the false defect area does not mean that the areas are defects, but needs further verification and analysis to determine whether the areas belong to the defect category.
Further, the processing of the seal area image information by the defect identifier, extracting a pseudo defect area, includes:
The method comprises the steps of obtaining a defect texture set, obtaining a positive sample sealing surface image of the biological product, wherein the positive sample sealing surface image is provided with a sealing area fold texture label, configuring defect compiling quantity constraint, randomly compiling the defect texture set to the positive sample sealing surface image according to the defect compiling quantity constraint, repeating the process for a plurality of times to obtain a negative sample sealing surface image, wherein the negative sample sealing surface image is provided with the defect texture label, outputting supervision data according to the positive sample sealing surface image and the negative sample sealing surface image serving as input data, and configuring the defect identifier by taking the defect texture label and the sealing area fold texture label as convolution network.
A series of defect texture images representing various possible seal area defects, such as breakage, smudge, fold anomalies, etc., are acquired from the big data and sorted into defect texture sets for subsequent compilation into a positive sample image. A series of biological sealing surface images are acquired from the big data as positive sample sealing surface images, and these positive sample images are labeled with sealing area fold texture labels to enable the normal texture to be distinguished from the defective texture during training, wherein the positive sample sealing surface images should have a clear sealing area and the sealing surface should be free of defects or have only normal fold texture. According to the practical requirement, configuring defect compiling quantity constraint, namely setting defect texture quantity compiled into positive sample images each time so as to control complexity and diversity of negative samples, randomly compiling textures in a defect texture set onto the positive sample sealing surface images by using an image processing technology, wherein each compiling follows the previously set defect compiling quantity constraint, repeating the compiling process for a plurality of times to generate enough negative sample sealing surface images, and labeling the negative sample images with defect texture labels so as to distinguish the negative sample images from the positive sample images in the training process. The method comprises the steps of using a positive sample sealing surface image and a negative sample sealing surface image as input data, using a defect texture label and a sealing area fold texture label as output supervision data to construct a Convolutional Neural Network (CNN), and obtaining a defect identifier after a model is subjected to repeated iterative training, wherein the defect identifier can accurately identify defect textures in a biological product sealing area.
And carrying out texture consistency analysis on the pseudo defect area to obtain a selected defect area.
After the processing of the image information of the sealing area is completed and the false defect area is extracted, the texture consistency analysis is carried out on the false defect area so as to further screen out the real defect area. The consistency and variability between different parts of the region are assessed by comparing their texture characteristics. If the majority of the texture features within the region are similar and conform to the texture pattern of the normal sealing surface, the region may be a false defect that is misjudged, whereas if there are significant texture variations or abnormal features within the region, it is more likely to be a truly defective region.
Further, as shown in fig. 2, performing texture consistency analysis on the pseudo defect area to obtain a selected defect area, including:
the method comprises the steps of carrying out longitudinal texture consistency analysis on the fake defect area to obtain a longitudinal selected defect area, carrying out transverse texture consistency analysis on the fake defect area to obtain a transverse selected defect area, and setting the union of the longitudinal selected defect area and the transverse selected defect area as the selected defect area.
Preferably, each of the pseudo-defect regions is subjected to texture uniformity analysis along its longitudinal direction (i.e., the direction perpendicular to the seal boundary) to obtain a longitudinal selected defect region, the pseudo-defect regions are subjected to texture uniformity analysis in the transverse direction (i.e., the direction parallel to the seal boundary) to obtain a transverse selected defect region by a method similar to the longitudinal texture uniformity analysis, and the union of the longitudinal selected defect region and the transverse selected defect region is set as the final selected defect region. Through the process, the texture consistency of the false defect area can be more comprehensively evaluated, and the real defect area can be more accurately identified.
Further, performing longitudinal texture consistency analysis on the pseudo defect region to obtain a longitudinal selected defect region, including:
The method comprises the steps of obtaining positioning coordinate information of a false defect area by falling the false defect area into a positioning space, traversing the false defect area based on the positioning coordinate information of the false defect area, extracting a feature set of the vertical distribution length of the false defect, carrying out outlier analysis according to the feature set of the vertical distribution length of the false defect to obtain a first outlier false defect, deleting the first outlier false defect from the positioning coordinate information of the false defect area to obtain positioning updating coordinate information of the false defect area, traversing adjacent false defect areas in the vertical direction according to the positioning updating coordinate information of the false defect area, extracting a feature set of the distribution interval of the false defect, wherein the defect interval is a defect vertical center distance, carrying out outlier analysis according to the feature set of the distribution interval of the false defect to obtain a second outlier false defect, and adding the first outlier false defect and the second outlier false defect into the longitudinally selected defect area.
The method comprises the steps of mapping each false defect area into a positioning space to obtain false defect area positioning coordinate information, traversing each false defect area to obtain the distribution length of the false defect area in the longitudinal direction (the direction perpendicular to a sealing boundary line) to form a vertical distribution length feature set of the false defects, conducting outlier analysis on the vertical distribution length feature set by using a statistical method (such as a box diagram and a Z-score), identifying outliers which are remarkably different from most data points, wherein the false defects corresponding to the outliers are regarded as first outlier false defects, deleting the first outlier false defects from the false defect area positioning coordinate information to obtain updated coordinate information, traversing the adjacent false defect areas in the longitudinal direction based on the updated false defect area positioning coordinate information to calculate the longitudinal center distance between the adjacent defects to serve as defect interval features, integrating all defect interval features to form the false defect distribution interval feature set, conducting outlier analysis on the defect interval feature set to identify outlier points which are remarkably different from other defects, and adding the outlier points which are remarkably different from other false defects into the first outlier defect and the second outlier defect area, and selecting the false defect to be significantly located in the second outlier defect. Through the above steps, the characteristics of the false defect region can be analyzed more finely, and outliers can be effectively identified and eliminated, so that the longitudinally selected defect region can be determined more accurately.
Further, performing outlier analysis according to the set of pseudo defect distribution interval features to obtain a second outlier pseudo defect, including:
And when the first distribution interval characteristic and the second distribution interval characteristic of the pseudo defect to be analyzed belong to the outlier interval characteristic set, adding the pseudo defect to be analyzed into the second outlier pseudo defect.
Preferably, the distribution pitch feature set of the false defects is subjected to an outlier analysis by using a statistical method (such as a box graph, an IQR, etc.), points which are significantly different from most pitch features are identified, pitch values corresponding to the points are regarded as outlier pitch features, all the outlier pitch features are integrated into one set, namely the outlier pitch feature set, a first distribution pitch feature and a second distribution pitch feature of the false defects to be analyzed are required to be calculated for each false defect to be analyzed, the first distribution pitch feature is the longitudinal center distance between the defect and the previous adjacent defect, the second distribution pitch feature is the longitudinal center distance between the defect and the next adjacent defect, the first distribution pitch feature and the second distribution pitch feature of the false defect to be analyzed are compared with the outlier pitch feature set, and if the two pitch features are all present in the outlier pitch feature set, namely that they are significantly deviated from the normal pitch range, the false defect to be analyzed can be considered to be abnormal in pitch, and likely to be a true defect, and the false defect to be added into the second distribution pitch feature set to be analyzed.
And when the size specification of the selected defect area meets the failure size specification, performing seal failure identification on the biological product.
After detailed analysis and verification of selected defect areas is completed, it can be assessed whether these defects have a substantial impact on the sealing performance of the bioproduct. Specifically, a series of failure dimensions are set according to the characteristics and the use requirements of the biological product, the dimensions define which defects in size and shape can cause sealing failure, the dimensions of the selected defect area are evaluated according to the preset failure dimensions, and if the dimensions of the selected defect area meet or exceed the preset failure dimensions, the biological product is immediately subjected to sealing failure identification.
Further, when the size specification of the selected defect area meets the failure size specification, performing seal failure identification on the biological product, including:
the method comprises the steps of extracting a color characteristic value set of a selected defect area, fitting the color characteristic value set through a defect color calibration network to obtain a defect color characteristic value set, distributing the defect color characteristic value set according to the color characteristic value distribution coordinate set to obtain a selected updated defect area, and carrying out seal failure identification on the biological product when the size specification of the selected updated defect area meets the failure size specification.
Preferably, the pixels in the selected defect area are subjected to color analysis, color characteristic values are extracted, the color characteristic values can comprise RGB values and HSV values, the color characteristic values are combined into a set, the distribution coordinates of each color characteristic value in the defect area are recorded at the same time to form a color characteristic value distribution coordinate set, a pre-trained defect color calibration network (such as a convolutional neural network CNN based on a deep learning model) is utilized to fit the extracted color characteristic value set, the defect color calibration network can identify a color characteristic mode highly related to the defect, a more accurate defect color characteristic value set is screened out from the original color characteristic value set, the screened defect color characteristic values are remapped back to an original image space to form a selected updated defect area, the position and shape of the actual defect are reflected more accurately, the size specification of the selected updated defect area is measured and comprises parameters such as length, width, area and the like, the size specification is compared with a preset size specification of the failure product, if any size specification of the selected updated defect area exceeds the preset size specification, the failure size specification is considered to be influenced by the biological size specification, and the failure performance of the biological product is considered to be influenced before the failure size is sealed according to the failure specification. Through the above process, the color features can be utilized to further refine and update the selected defect region, thereby improving the accuracy and reliability of seal detection.
Further, fitting the color eigenvalue set through a defect color calibration network to obtain a defect color eigenvalue set, including:
Collecting a non-sealing failure defect color characteristic value set and a sealing failure defect color characteristic value set, constructing a classifier input data set, and configuring the defect color calibration network according to the classifier input data set.
Selecting the areas containing various non-sealing failure defects (such as scratches, stains, chromatic aberration and the like) from the non-sealing failure biological product samples, extracting color characteristic values from the areas, combining the color characteristic values and labels corresponding to the color characteristic values (non-sealing failure) to form a non-sealing failure defect color characteristic value set as a part of a classifier input data set, selecting the areas containing the sealing failure defects (such as holes, cracks, bubbles and the like) from the sealing failure biological product samples, extracting the color characteristic values from the areas, combining the color characteristic values and labels corresponding to the color characteristic values (sealing failure) to form a sealing failure defect color characteristic value set as another part of the classifier input data set, combining the non-sealing failure defect color characteristic value set and the sealing failure defect color characteristic value set to form a complete classifier input data set, and training a defect color calibration network by using the constructed classifier input data set, wherein the defect color calibration network is constructed based on a convolutional neural network and can output a model of the defect type (non-sealing failure or sealing failure) according to the color characteristic value input. Through the steps, a defect color calibration network based on color characteristic values can be constructed and used for identifying seal failure defects in biological products.
In summary, the embodiment of the application has at least the following technical effects:
Firstly, sealing a biological product through a first flexible sealing material and a second flexible sealing material and transmitting the biological product to a first detection platform, wherein the biological product is provided with a sealing area positioning coordinate on the first detection platform, and the sealing area is a sealing interface area of the first flexible sealing material and the second flexible sealing material. Then, the camera is activated, and sealing area image information of the biological product is acquired based on the sealing area positioning coordinates. Then, the seal region image information is processed by a defect identifier to extract a pseudo defect region. And finally, carrying out texture consistency analysis on the pseudo defect area to obtain a selected defect area, and carrying out seal failure identification on the biological product when the size specification of the selected defect area meets the failure size specification. The technical problem of biological product seal failure detection accuracy is relatively poor in the prior art is solved, and the technical effects of improving detection efficiency and accuracy are achieved through an image recognition technology.
It should be noted that the sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. And the foregoing description has been directed to specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
The foregoing description of the preferred embodiments of the application is not intended to limit the application to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and scope of the application are intended to be included within the scope of the application.
The specification and figures are merely exemplary illustrations of the present application and are considered to cover any and all modifications, variations, combinations, or equivalents that fall within the scope of the application. It will be apparent to those skilled in the art that various modifications and variations can be made to the present application without departing from the scope of the application. Thus, the present application is intended to include such modifications and alterations insofar as they come within the scope of the application or the equivalents thereof.
Claims (4)
1. The biological product sealing failure detection method based on image recognition is characterized by comprising the following steps of:
Sealing a biological product through a first flexible sealing material and a second flexible sealing material and transmitting the biological product to a first detection platform, wherein the biological product is provided with a sealing area positioning coordinate on the first detection platform, and the sealing area is a sealing junction area of the first flexible sealing material and the second flexible sealing material;
activating a camera, and acquiring sealing area image information of the biological product based on the positioning coordinates of the sealing area;
processing the image information of the sealing area through a defect identifier, and extracting a false defect area;
the pseudo defect region refers to a region having a significant difference from a normal region;
performing texture consistency analysis on the pseudo defect area to obtain a selected defect area;
When the size specification of the selected defect area meets the failure size specification, performing seal failure identification on the biological product;
Processing the seal region image information by a defect identifier to extract a pseudo defect region, including:
Obtaining a defect texture set;
obtaining a positive specimen sealing surface image of the biological product, wherein the positive specimen sealing surface image has a sealing region crumpled texture label;
configuring defect compiling quantity constraint;
Randomly compiling the defect texture set into the positive sample sealing surface image according to the defect compiling number constraint, and repeating for a plurality of times to obtain a negative sample sealing surface image, wherein the negative sample sealing surface image is provided with a defect texture label;
According to the positive sample sealing surface image and the negative sample sealing surface image which are used as input data, outputting supervision data by taking the defect texture label and the sealing area fold texture label as convolution networks, and configuring the defect identifier;
the false defect area is subjected to texture consistency analysis to obtain a selected defect area, and the method comprises the following steps:
performing longitudinal texture consistency analysis on the pseudo defect area to obtain a longitudinal selected defect area;
performing transverse texture consistency analysis on the pseudo defect area to obtain a transverse selected defect area;
Setting the union of the longitudinally selected defect area and the transversely selected defect area as the selected defect area;
the false defect area is subjected to longitudinal texture consistency analysis to obtain a longitudinal selected defect area, and the method comprises the following steps:
the false defect area falls into a positioning space to obtain positioning coordinate information of the false defect area;
Traversing the false defect area based on the positioning coordinate information of the false defect area, and extracting a feature set of the longitudinal distribution length of the false defect;
performing outlier analysis according to the longitudinal distribution length feature set of the false defects to obtain first outlier false defects;
Deleting the first outlier pseudo defect from the pseudo defect area positioning coordinate information to obtain pseudo defect area positioning updating coordinate information;
According to the positioning updating coordinate information of the false defect areas, traversing the adjacent false defect areas in the longitudinal direction, and extracting a false defect distribution interval characteristic set, wherein the defect interval is the longitudinal center distance of the defects;
performing outlier analysis according to the false defect distribution interval feature set to obtain a second outlier false defect;
Adding the first and second outlier pseudo defects into the longitudinally selected defect region;
The false defect distribution interval feature set performs outlier analysis to obtain a second outlier false defect, including:
performing outlier analysis according to the false defect distribution interval feature set to obtain an outlier interval feature set;
when the first distribution interval characteristic and the second distribution interval characteristic of the to-be-analyzed pseudo defect belong to the outlier interval characteristic set, adding the to-be-analyzed pseudo defect into the second outlier pseudo defect;
The first distribution interval characteristic refers to the longitudinal center distance between the false defect to be analyzed and the previous adjacent defect;
the second distribution interval characteristic refers to the longitudinal center distance between the false defect to be analyzed and the next adjacent defect.
2. The method for detecting seal failure of a biological product based on image recognition according to claim 1, wherein the biological product is sealed by a first flexible sealing material and a second flexible sealing material and transferred to a first detection platform, and the biological product has a sealing area positioning coordinate on the first detection platform, comprising:
Obtaining a seal control parameter of the first flexible sealing material and the second flexible sealing material on the biological product, wherein the seal control parameter has a coordinate feature identifier;
performing regional cluster analysis on the coordinate feature identifiers according to the seal control parameters to obtain regional cluster results, wherein the same category of the regional cluster results has the same seal control parameters;
And adding the region clustering result into the sealing region positioning coordinates.
3. The method of image recognition-based seal failure detection of a biological product of claim 1, wherein when the size specification of the selected defect region meets a failure size specification, seal failure identification of the biological product comprises:
Extracting a color characteristic value set of the selected defect area, wherein the color characteristic value set is provided with a color characteristic value distribution coordinate set;
fitting the color characteristic value set through a defect color calibration network to obtain a defect color characteristic value set;
Distributing the defect color characteristic value set according to the color characteristic value distribution coordinate set to obtain a selected updated defect region;
And when the size specification of the selected updated defect area meets the failure size specification, performing seal failure identification on the biological product.
4. The method for detecting seal failure of biological product based on image recognition according to claim 3, wherein fitting the set of color eigenvalues through a defect color calibration network to obtain a set of defect color eigenvalues comprises:
Collecting a non-sealing failure defect color characteristic value set and a sealing failure defect color characteristic value set, and constructing a classifier input data set;
and configuring the defect color calibration network according to the classifier input data set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411462732.2A CN118961756B (en) | 2024-10-19 | 2024-10-19 | A method for detecting sealing failure of biological products based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411462732.2A CN118961756B (en) | 2024-10-19 | 2024-10-19 | A method for detecting sealing failure of biological products based on image recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118961756A CN118961756A (en) | 2024-11-15 |
| CN118961756B true CN118961756B (en) | 2025-01-21 |
Family
ID=93402710
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411462732.2A Active CN118961756B (en) | 2024-10-19 | 2024-10-19 | A method for detecting sealing failure of biological products based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118961756B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102663391A (en) * | 2012-02-27 | 2012-09-12 | 安科智慧城市技术（中国）有限公司 | Image multifeature extraction and fusion method and system |
| CN114120318A (en) * | 2021-11-29 | 2022-03-01 | 哈尔滨工业大学 | Dark field image target point accurate extraction method based on integrated decision tree |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5898491A (en) * | 1997-03-28 | 1999-04-27 | Hitachi Electronics Engineering Co. Ltd. | Surface defect test method and surface defect tester |
| CN106127780B (en) * | 2016-06-28 | 2019-01-18 | 华南理工大学 | A kind of curved surface defect automatic testing method and its device |
| CN111161246B (en) * | 2019-12-30 | 2024-05-14 | 歌尔股份有限公司 | Product defect detection method, device and system |
| CN115457037B (en) * | 2022-11-10 | 2023-03-24 | 澳润(山东)药业有限公司 | Raw material quality inspection method for donkey-hide gelatin product production |
| CN115797354B (en) * | 2023-02-09 | 2023-05-30 | 厦门微亚智能科技有限公司 | Method for detecting appearance defects of laser welding seam |
| CN116559183B (en) * | 2023-07-11 | 2023-11-03 | 钛玛科（北京）工业科技有限公司 | Method and system for improving defect judging efficiency |
| CN118334015B (en) * | 2024-06-12 | 2024-08-16 | 数智汇能(大连)科技发展有限公司 | Defect recognition method, system, equipment and medium based on visual image |
-
2024
- 2024-10-19 CN CN202411462732.2A patent/CN118961756B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102663391A (en) * | 2012-02-27 | 2012-09-12 | 安科智慧城市技术（中国）有限公司 | Image multifeature extraction and fusion method and system |
| CN114120318A (en) * | 2021-11-29 | 2022-03-01 | 哈尔滨工业大学 | Dark field image target point accurate extraction method based on integrated decision tree |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118961756A (en) | 2024-11-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR102749767B1 (en) | Detecting defects in semiconductor specimens using weak labeling | |
| CN110136101B (en) | Tire X-ray defect detection method based on twinning distance comparison | |
| CN113592828B (en) | Nondestructive testing method and system based on industrial endoscope | |
| Eshkevari et al. | Automatic dimensional defect detection for glass vials based on machine vision: A heuristic segmentation method | |
| CN117392042A (en) | Defect detection method, defect detection equipment and storage medium | |
| CN111861966B (en) | Model training method and device and defect detection method and device | |
| CN113643268A (en) | Industrial product defect quality inspection method and device based on deep learning and storage medium | |
| JP5718781B2 (en) | Image classification apparatus and image classification method | |
| CN110796661B (en) | Fungal microscopic image segmentation detection method and system based on convolutional neural network | |
| CN110648310A (en) | Weakly Supervised Casting Defect Recognition Method Based on Attention Mechanism | |
| CN111652883B (en) | Glass surface defect detection method based on deep learning | |
| CN114037679B (en) | A product image defect detection method and device based on unsupervised feature combination | |
| CN117169282B (en) | Quality detection method and system of composite film for medicine packaging | |
| CN115035092A (en) | Image-based bottle detection method, device, equipment and storage medium | |
| CN115294323A (en) | Foreign matter detection method based on positive sample and storage medium | |
| JPH10209230A (en) | Failure analysis apparatus and method | |
| CN111753877B (en) | Product quality detection method based on deep neural network migration learning | |
| CN113109240A (en) | Method and system for determining imperfect grains of grains implemented by computer | |
| CN112419244B (en) | Concrete crack segmentation method and device | |
| CN113744252A (en) | Method, apparatus, storage medium and program product for marking and detecting defects | |
| CN115937555A (en) | An Industrial Defect Detection Algorithm Based on Standardized Flow Model | |
| CN118961756B (en) | A method for detecting sealing failure of biological products based on image recognition | |
| CN119293735B (en) | Method, system, device and medium for automatically determining pallet defects | |
| Shin et al. | isualization for Explanation of Deep Learning-Based Defect Detection Model Using Class Activation Map. | |
| CN119919410A (en) | Method for measuring size and appearance of O-rings |
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 |