CN114757945A - In-situ identification method and device for corrosion and coating aging of metal matrix - Google Patents
In-situ identification method and device for corrosion and coating aging of metal matrix Download PDFInfo
- Publication number
- CN114757945A CN114757945A CN202210661350.7A CN202210661350A CN114757945A CN 114757945 A CN114757945 A CN 114757945A CN 202210661350 A CN202210661350 A CN 202210661350A CN 114757945 A CN114757945 A CN 114757945A
- Authority
- CN
- China
- Prior art keywords
- corrosion
- identification
- coating
- model
- convolution
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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]
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30136—Metal
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Quality & Reliability (AREA)
- Testing Resistance To Weather, Investigating Materials By Mechanical Methods (AREA)
Abstract
The invention provides an in-situ identification method and device for corrosion of a metal substrate and aging of a coating. The method comprises the following steps: segmenting the damage image according to a preset target image segmentation model to obtain a plurality of segmentation subimages; determining prior frame sizes and category information of prior frames of a plurality of segmented sub-images; constructing an initial matrix corrosion identification model and an initial coating aging identification model, training the models based on segmentation subimages, prior frame sizes and category information, and obtaining a target matrix corrosion identification model and a target coating aging identification model; acquiring an image to be identified on the metal surface, and inputting the image to be identified into a target image segmentation model to obtain a plurality of sub-images to be identified; and identifying a plurality of sub-images to be identified respectively based on the target matrix corrosion identification model and the target coating aging identification model to obtain a matrix corrosion identification result and a coating aging identification result. The invention can improve the identification precision for identifying the corrosion of the metal matrix and the aging of the coating.
Description
Technical Field
The invention relates to the technical field of image recognition, in particular to an in-situ recognition method and device for corrosion and coating aging of a metal matrix.
Background
The high-end equipment metal material in the fields of sea, land, air and the like is used in high-temperature, high-humidity, high-salt-fog, strong solar radiation and frequent dry/wet alternate action environments for a long time, and easily causes surface fine damage defects such as corrosion of a metal matrix, aging of a coating and the like, so that a key structural part and a part of functional parts of the equipment lose effectiveness in advance. For example, 7A85 aluminum alloy as an important bearing structural material of a novel airplane in China can encounter a plurality of harsh use environments such as ocean atmosphere, dry and hot desert and the like in the service process, and factors such as humidity, corrosive media and the like in the atmospheric environment tend to cause corrosive influence on the alloy; the 7A52 high-strength aluminum alloy welding piece is extremely sensitive to stress corrosion cracking, and particularly, the welding piece inevitably has metallurgical defects and residual stress distribution, so that the stress corrosion failure tendency of the welding piece is greatly enhanced. These apparent minor damages not only directly affect the operation safety of the equipment, but also bring huge economic burden to the inspection and maintenance work of the equipment. According to the international air transport association report, the cost of regular maintenance of the aircraft and replacement of structural parts due to corrosion is $ 10-20 per hour. Kovich academy of metal research institute of china academy of sciences: the economic loss caused by corrosion in China accounts for about 5 percent of the total value of national production every year.
The appearance damage fine characteristic image of the material is one of important characteristics for qualitatively and quantitatively evaluating the corrosion and coating aging performance of a metal material matrix. In recent years, the application of image recognition technology and deep learning in the field of corrosion science has achieved a great deal of research results, but the following problems still exist: the network model in the prior art is not suitable for the detection of a small target of apparent damage (the damaged area is less than 1% of the surface area or the apparent tiny damaged target with the size of less than 20 multiplied by 20 pixels), and has low identification precision on the corrosion of a metal matrix and the aging of a coating on the metal surface.
Therefore, it is urgently needed to provide an in-situ identification method and device for metal matrix corrosion and coating aging, so as to solve the technical problem that the identification precision for metal matrix corrosion and coating aging on a metal surface is low in the prior art.
Disclosure of Invention
In view of this, it is necessary to provide an in-situ identification method and device for metal substrate corrosion and coating aging, so as to solve the technical problem in the prior art that the monitoring accuracy for detecting the refrigerator food material is low.
In one aspect, the invention provides a method for in-situ identification of corrosion and coating aging of a metal substrate, comprising:
Acquiring a damage image of the metal surface, and segmenting the damage image according to a preset target image segmentation model to obtain a plurality of segmentation sub-images;
determining the prior frame size of the prior frame of the plurality of segmentation sub-images based on a preset clustering method, and determining the category information of the prior frame;
constructing an initial matrix corrosion identification model, and training the initial matrix corrosion identification model based on the plurality of segmentation sub-images, the prior frame size and the category information to obtain a target matrix corrosion identification model with complete training;
constructing an initial coating aging identification model, and training the initial coating aging identification model based on the plurality of segmentation sub-images, the prior frame size and the category information to obtain a target coating aging identification model;
acquiring an image to be identified on the surface of a metal, and inputting the image to be identified into the target image segmentation model to obtain a plurality of sub-images to be identified;
and identifying the plurality of sub-images to be identified respectively based on the target matrix corrosion identification model and the target coating aging identification model, and correspondingly obtaining a matrix corrosion identification result and a coating aging identification result.
In some possible implementation manners, after the segmenting the damage image according to a preset target image segmentation model to obtain a plurality of segmentation sub-images, the method further includes:
acquiring a plurality of real sub-images of the damage image, and determining the intersection ratio of the plurality of real sub-images and the plurality of segmentation sub-images;
and determining the segmentation accuracy of the target image segmentation model according to the intersection ratio.
In some possible implementations, the target image segmentation model includes a coding unit and a decoding unit, the coding unit includes a depth convolution module, a multi-feature extraction module, and a first convolution layer; the decoding unit comprises a second convolution layer, a first up-sampling layer, a superposition layer, a third convolution layer and a second up-sampling layer;
the depth convolution module is used for extracting low-dimensional features and high-dimensional features of the damage image;
the multi-feature extraction module is used for receiving the high-dimensional features and generating combined features;
the first convolution layer is used for performing convolution processing on the combined features to generate first convolution features;
the second convolution layer is used for carrying out convolution processing on the low-dimensional features to generate second convolution features;
The first upsampling layer is used for upsampling the first convolution feature to obtain a first upsampled feature;
the superposition layer is used for superposing the first upsampling feature and the second convolution feature to obtain a superposition feature;
the third convolution layer is used for carrying out convolution processing on the superposition characteristics to generate third convolution characteristics;
the second upsampling layer is configured to upsample the third convolution feature to obtain the multiple segmentation sub-images.
In some possible implementations, the multi-feature extraction module includes a first sub-convolutional layer, a first hole convolutional layer, a second hole convolutional layer, a third hole convolutional layer, a pooling layer, and a stitching layer in parallel;
the convolution kernel of the first sub convolution layer is 1 multiplied by 1 and is used for extracting first feature information of the high-dimensional features;
the convolution kernel of the first cavity convolution layer is 3 x 3, the sampling rate is 6, and the second feature information of the high-dimensional features is extracted;
the convolution kernel of the second void convolution layer is 3 x 3, the sampling rate is 12, and the second void convolution layer is used for extracting third feature information of the high-dimensional features;
the convolution kernel of the third void convolution layer is 3 x 3, the sampling rate is 18, and the fourth feature information of the high-dimensional features is extracted;
The pooling layer is used for extracting fifth feature information of the high-dimensional features;
the splicing layer is used for splicing the first feature information, the second feature information, the third feature information, the fourth feature information and the fifth feature information to generate the combined feature.
In some possible implementations, the substrate corrosion identification result includes a substrate corrosion category including pitting and cracking, and a substrate corrosion parameter including a number of pitting, a pitting aperture, a pitting depth, a breach size, and a cracking depth; the coating aging identification result comprises a coating damage type, wherein the coating damage type comprises mildew, chalking, cracking and falling.
In some possible implementations, the in-situ identification method of metal substrate corrosion and coating aging further includes:
constructing the correlation relationship between the erosion hole number, the erosion hole aperture, the erosion hole depth, the split size and the cracking depth and the corrosion damage degree of the matrix;
and determining the corrosion damage degree of the substrate on the metal surface according to the number of the etching holes, the aperture of the etching holes, the depth of the etching holes, the size of the gap, the cracking depth and the correlation relationship.
In some possible implementations, the coating degradation identification result further includes a coating HSV color space; the in-situ identification method for the corrosion and the coating aging of the metal substrate further comprises the following steps:
determining similarity of the coating HSV color space and a standard HSV color space;
and determining the coating aging degree of the metal surface according to the similarity.
In some possible implementations, after the acquiring the damage image of the metal surface, the method further includes:
and constructing a filtering window, and carrying out filtering processing on the damaged image based on the filtering window to obtain a filtering image.
In some possible implementations, the model structure of the target substrate corrosion identification model is a YOLOv3 model structure, and the model structure of the target coating aging identification model is an SSD model structure.
In another aspect, the present invention further provides an in-situ identification apparatus for corrosion and coating aging of a metal substrate, comprising:
the image segmentation module is used for acquiring a damage image of the metal surface, segmenting the damage image according to a preset target image segmentation model and acquiring a plurality of segmentation sub-images;
the prior frame determining module is used for determining prior frame sizes of prior frames of the plurality of segmentation sub-images based on a preset clustering method and determining category information of the prior frames;
The corrosion identification model building module is used for building an initial matrix corrosion identification model, and training the initial matrix corrosion identification model based on the plurality of segmentation sub-images, the prior frame size and the category information to obtain a target matrix corrosion identification model with complete training;
the coating aging identification model building module is used for building an initial coating aging identification model, and training the initial coating aging identification model based on the plurality of segmentation sub-images, the prior frame size and the category information to obtain a target coating aging identification model;
the to-be-identified image acquisition module is used for acquiring an to-be-identified image of the metal surface, inputting the to-be-identified image into the target image segmentation model and acquiring a plurality of to-be-identified sub-images;
and the identification module is used for identifying the plurality of sub-images to be identified respectively based on the target matrix corrosion identification model and the target coating aging identification model, and correspondingly obtaining a matrix corrosion identification result and a coating aging identification result.
The beneficial effects of adopting the above embodiment are: according to the in-situ identification method for metal matrix corrosion and coating aging, the priori frame sizes of the priori frames of the plurality of segmentation sub-images are determined based on the preset clustering method, so that the priori frames can be more attached to the identification scenes of metal matrix corrosion and coating aging, and the identification accuracy for identifying the metal matrix corrosion and the coating aging can be improved.
Furthermore, the method and the device can further improve the identification precision of identifying the substrate corrosion and the coating aging of the metal surface by identifying a plurality of sub-images to be identified respectively based on the target substrate corrosion identification model and the target coating aging identification model and correspondingly obtaining the substrate corrosion identification result and the coating aging identification result instead of identifying the substrate corrosion and the coating aging through one model.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the description below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained based on these drawings without creative efforts.
FIG. 1 is a schematic flow chart of one embodiment of a method for in-situ identification of corrosion and coating degradation of a metal substrate according to the present invention;
FIG. 2 is a flowchart illustrating an embodiment of determining segmentation accuracy provided by the present invention;
FIG. 3 is a schematic structural diagram of an embodiment of a target image segmentation model provided in the present invention;
FIG. 4 is a schematic flow chart illustrating one embodiment of the present invention for determining the degree of corrosion damage to a substrate;
FIG. 5 is a schematic flow chart of one embodiment of the method for determining the degree of aging damage of a coating according to the present invention;
FIG. 6 is a schematic structural diagram of an in-situ identification device for corrosion and coating aging of a metal substrate according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of an embodiment of an electronic device provided in 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. It should be apparent that the described embodiments are only some embodiments of the present invention, and not all 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.
It should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this invention illustrate operations performed in accordance with some embodiments of the present invention. It should be understood that the operations of the flow diagrams may be performed out of order, and that steps without logical context may be performed in reverse order or concurrently. One skilled in the art, under the direction of this summary, may add one or more other operations to, or remove one or more operations from, the flowchart.
Some of the block diagrams shown in the figures are functional entities and do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different networks and/or processor systems and/or microcontroller systems.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the invention. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the embodiments described herein can be combined with other embodiments.
The embodiment of the invention provides an in-situ identification method and device for corrosion of a metal substrate and aging of a coating, which are respectively explained below.
Fig. 1 is a schematic flow chart of an embodiment of the in-situ identification method for corrosion and coating aging of a metal substrate provided by the present invention, as shown in fig. 1, the in-situ identification method for corrosion and coating aging of a metal substrate includes:
S101, obtaining a damage image of the metal surface, and segmenting the damage image according to a preset target image segmentation model to obtain a plurality of segmentation sub-images;
s102, determining prior frame sizes of prior frames of a plurality of segmented sub-images based on a preset clustering method, and determining category information of the prior frames;
s103, constructing an initial matrix corrosion identification model, and training the initial matrix corrosion identification model based on a plurality of segmentation sub-images, the prior frame size and the category information to obtain a target matrix corrosion identification model with complete training;
s104, constructing an initial coating aging identification model, and training the initial coating aging identification model based on a plurality of segmentation sub-images, the prior frame size and the category information to obtain a target coating aging identification model;
s105, obtaining an image to be recognized on the metal surface, inputting the image to be recognized into a target image segmentation model, and obtaining a plurality of sub-images to be recognized;
s106, identifying a plurality of sub-images to be identified respectively based on the target matrix corrosion identification model and the target coating aging identification model, and correspondingly obtaining a matrix corrosion identification result and a coating aging identification result.
Compared with the prior art, the in-situ identification method for metal matrix corrosion and coating aging provided by the embodiment of the invention has the advantages that the prior frame size of the prior frame of the plurality of segmentation sub-images is determined based on the preset clustering method, so that the prior frame can be more attached to the identification scene of metal matrix corrosion and coating aging, and the identification precision of identifying the metal matrix corrosion and coating aging can be improved.
Furthermore, the embodiment of the invention identifies a plurality of sub-images to be identified respectively based on the target matrix corrosion identification model and the target coating aging identification model, and correspondingly obtains the matrix corrosion identification result and the coating aging identification result, rather than identifying the matrix corrosion and the coating aging through one model, so that the identification accuracy of identifying the matrix corrosion and the coating aging of the metal surface can be further improved.
In some embodiments of the present invention, if the clustering method in step S102 is a K-means clustering method, step S102 specifically includes:
defining k as a clustering number, randomly selecting k points in the damage image to be recorded as a central point, and calculating the Euclidean distance d from each pixel point to the central point in the damage image, wherein d is as follows:
wherein (x)1,y1) The coordinate value of the pixel point; (x)2,y2) Are coordinate values of the center point.
Determining k clustering clusters according to the Euclidean distance, and calculating a preselected central point of each clustering cluster;
recalculating each pixel point to obtain the distance between the preselected central points to form k new cluster points, calculating the central points to be compared of the new cluster points, judging whether the difference value between the preselected central points and the central points to be compared is smaller than a preset difference value, if so, determining the prior frame size according to the new cluster points, and if so, re-determining the cluster points.
Since the metal surface of the high-end equipment is large and complex, in order to avoid missing detection, in some embodiments of the present invention, the step S101 of obtaining the loss image of the metal surface specifically includes:
the method comprises the steps of obtaining a damage image of a metal surface based on the intelligent robot, and planning an acquisition path of the intelligent robot before obtaining the damage image.
By planning the acquisition path of the intelligent robot, the integrity and the speed of the acquired metal surface of the high-end equipment can be improved, so that the in-situ identification efficiency and the accuracy of metal matrix corrosion and coating aging can be further improved.
In order to avoid that when the segmentation accuracy of the target image segmentation model is low, the identification of the subsequent metal surface for substrate corrosion and coating aging is affected, which results in low identification accuracy, in some embodiments of the present invention, as shown in fig. 2, after step S101, the method further includes:
s201, obtaining a plurality of real sub-images of the damage image, and determining the intersection ratio of the plurality of real sub-images and the plurality of segmentation sub-images;
and S202, determining the segmentation accuracy of the target image segmentation model according to the intersection ratio.
According to the method, the segmentation accuracy of the target image segmentation model is determined, and when the segmentation accuracy meets the requirement, the steps S102-S106 are performed, so that the identification accuracy of metal matrix corrosion and coating aging is further improved.
Specifically, when the intersection ratio is greater than or equal to 0.5, the segmentation accuracy of the target image segmentation model meets the requirement.
In an embodiment of the present invention, as shown in fig. 3, the target image segmentation model includes a coding unit and a decoding unit, where the coding unit includes a depth convolution module, a multi-feature extraction module, and a first convolution layer; the decoding unit comprises a second convolution layer, a first up-sampling layer, a superposition layer, a third convolution layer and a second up-sampling layer;
the depth convolution module is used for extracting low-dimensional features and high-dimensional features of the damaged image;
the multi-feature extraction module is used for receiving the high-dimensional features and generating combined features;
the first convolution layer is used for performing convolution processing on the combined features to generate first convolution features;
the second convolution layer is used for carrying out convolution processing on the low-dimensional features to generate second convolution features;
the first upsampling layer is used for upsampling the first convolution characteristic to obtain a first upsampled characteristic;
the superposition layer is used for superposing the first up-sampling feature and the second convolution feature to obtain a superposition feature;
the third convolution layer is used for carrying out convolution processing on the superposition characteristics to generate third convolution characteristics;
the second upsampling layer is used for upsampling the third convolution characteristic to obtain a plurality of segmentation sub-images.
Specifically, the method comprises the following steps: the multi-feature extraction module comprises a first sub-convolution layer, a first cavity convolution layer, a second cavity convolution layer, a third cavity convolution layer, a pooling layer and a splicing layer which are parallel;
the convolution kernel of the first sub-convolution layer is 1 multiplied by 1 and is used for extracting first characteristic information of high-dimensional characteristics;
the convolution kernel of the first cavity convolution layer is 3 multiplied by 3, the sampling rate is 6, and the second characteristic information of the high-dimensional characteristic is extracted;
the convolution kernel of the second cavity convolution layer is 3 multiplied by 3, the sampling rate is 12, and the second cavity convolution layer is used for extracting third feature information of high-dimensional features;
the convolution kernel of the third void convolution layer is 3 multiplied by 3, the sampling rate is 18, and the fourth feature information of the high-dimensional features is extracted;
the pooling layer is used for extracting fifth feature information of the high-dimensional features;
the splicing layer is used for splicing the first characteristic information, the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information to generate combined characteristics.
According to the embodiment of the invention, the high-dimensional characteristics are extracted by setting the cavity convolution layers with different sampling rates, and the receptive field information with different scales can be captured, so that the characteristic information with different scales is captured, and the segmentation precision is improved.
In a specific embodiment of the invention, the substrate corrosion identification result comprises a substrate corrosion category and substrate corrosion parameters, the substrate corrosion category comprises pitting corrosion and cracking, and the substrate corrosion parameters comprise the number of etching holes, the aperture of the etching holes, the depth of the etching holes, the size of a crack and the depth of the crack; the coating aging identification result comprises coating damage types including mildew, chalking, cracks and falling.
The embodiment of the invention not only can obtain the corrosion category of the base body, but also can obtain the specific corrosion parameters of the base body, thereby providing more identification information for users.
In order to overcome the defect that a user cannot intuitively know the corrosion damage degree of the metal substrate due to a large number of recognition results, in some embodiments of the present invention, as shown in fig. 4, the in-situ recognition method for metal substrate corrosion and coating aging further includes:
s401, constructing the correlation between the number of etching holes, the aperture of the etching holes, the depth of the etching holes, the size of a crack and the corrosion damage degree of a matrix;
s402, determining the corrosion damage degree of the substrate on the metal surface according to the number of the etching holes, the hole diameter of the etching holes, the depth of the etching holes, the size of the crack, the cracking depth and the correlation.
In the embodiment of the present invention, the pitting damage degree is classified into 5 grades, respectively, a slight, a small, a medium, a large and a serious, according to the number of etching holes, the hole diameter of the etching holes and the depth of the etching holes per unit area, and similarly, the cracking damage grade is also classified into 5 grades, respectively, a slight, a small, a medium, a large and a serious, according to the size of the crack and the depth of the crack.
Through the setting, a user can visually obtain the corrosion damage degree of the base body according to the damage grade, and take corresponding measures according to the corrosion damage degree of the base body, so that the response speed to the corrosion damage is improved.
Further, since the color of the metal surface changes due to the aging of the coating, in order to facilitate the user to intuitively obtain the aging degree of the coating, in some embodiments of the present invention, the coating aging recognition result further includes a coating HSV color space; then as shown in fig. 5, the method for identifying the corrosion of the metal substrate and the aging of the coating in situ further comprises:
s501, determining the similarity between the coating HSV color space and the standard HSV color space;
and S502, determining the coating aging degree of the metal surface according to the similarity.
Wherein, the HSV color space includes three color parameters, which are hue (H), saturation (S), and lightness (V), and the similarity is:
in the formula (I), the compound is shown in the specification,
is the similarity; (H)i ,Si ,Vi) Is the coating HSV color space; (H)0 ,S0 ,V0) Is a standard HSV color space.
It should be noted that:
closer to 1, indicates less aging of the coating, and conversely,
the closer to 0, the greater the degree of aging of the coating.
In order to avoid the low recognition accuracy caused by the excessive noise in the damage image, in some embodiments of the present invention, after acquiring the damage image of the metal surface, the method further includes:
and constructing a filtering window, and carrying out filtering processing on the damaged image based on the filtering window to obtain a filtering image.
Specifically, the method comprises the following steps: and constructing a3 × 3 filtering window to filter each pixel point in the damage image, wherein if the pixel values in the 3 × 3 filtering window are a1, a2, a3, a4, a5, a6, a7, a8 and a9, the nine pixel values are sorted, and the median value is selected as the pixel value of the current pixel point.
According to the embodiment of the invention, the influence of noise in the damaged image on the identification precision can be reduced by filtering the damaged image, and the identification precision of the in-situ identification method for metal matrix corrosion and coating aging is further improved.
Furthermore, the damage image can be subjected to illumination uniformization processing, and the influence of image noise on the identification precision is further eliminated.
In an embodiment of the invention, the model structure of the target substrate corrosion identification model is a Yolov3 model structure, and the model structure of the target coating aging identification model is an SSD model structure.
In order to better implement the in-situ identification method for corrosion and coating aging of a metal substrate in the embodiment of the present invention, on the basis of the in-situ identification method for corrosion and coating aging of a metal substrate, correspondingly, the embodiment of the present invention also provides an in-situ identification device for corrosion and coating aging of a metal substrate, wherein multiple data sources include a first data source and at least one second data source, as shown in fig. 6, the in-situ identification device 600 for corrosion and coating aging of a metal substrate includes:
The image segmentation module 601 is configured to obtain a damage image of a metal surface, and segment the damage image according to a preset target image segmentation model to obtain a plurality of segmentation sub-images;
a prior frame determining module 602, configured to determine prior frame sizes of prior frames of the multiple segmented sub-images based on a preset clustering method, and determine category information of the prior frames;
the corrosion identification model construction module 603 is used for constructing an initial matrix corrosion identification model, and training the initial matrix corrosion identification model based on a plurality of segmentation subimages, the prior frame size and the category information to obtain a target matrix corrosion identification model with complete training;
the coating aging identification model building module 604 is used for building an initial coating aging identification model, and training the initial coating aging identification model based on a plurality of segmentation subimages, the prior frame size and the category information to obtain a target coating aging identification model;
the to-be-identified image acquisition module 605 is configured to acquire an to-be-identified image of a metal surface, and input the to-be-identified image into the target image segmentation model to obtain a plurality of to-be-identified sub-images;
the identification module 606 is configured to identify a plurality of sub-images to be identified based on the target matrix corrosion identification model and the target coating aging identification model, and correspondingly obtain a matrix corrosion identification result and a coating aging identification result.
The in-situ recognition apparatus 600 for metal substrate corrosion and coating aging provided in the foregoing embodiment may implement the technical solutions described in the foregoing in-situ recognition method embodiments for metal substrate corrosion and coating aging, and the specific implementation principles of each module or unit may refer to the corresponding contents in the foregoing in-situ recognition method embodiments for metal substrate corrosion and coating aging, which are not described herein again.
As shown in fig. 7, the present invention further provides an electronic device 700. The electronic device 700 includes a processor 701, a memory 702, and a display 703. Fig. 7 shows only some of the components of the electronic device 700, but it is to be understood that not all of the shown components are required to be implemented, and that more or fewer components may be implemented instead.
The processor 701 may be, in some embodiments, a Central Processing Unit (CPU), microprocessor or other data Processing chip for executing program codes stored in the memory 702 or Processing data, such as in-situ identification of metal substrate corrosion and coating aging in the present invention.
In some embodiments, processor 701 may be a single server or a group of servers. The server groups may be centralized or distributed. In some embodiments, the processor 701 may be local or remote. In some embodiments, processor 701 may be implemented in a cloud platform. In an embodiment, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an intra-site, a multi-cloud, and the like, or any combination thereof.
The storage 702 may in some embodiments be an internal storage unit of the electronic device 700, such as a hard disk or a memory of the electronic device 700. The memory 702 may also be an external storage device of the electronic device 700 in other embodiments, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), etc., provided on the electronic device 700.
Further, the memory 702 may also include both internal storage units and external storage devices of the electronic device 700. The memory 702 is used for storing application software and various types of data for installing the electronic apparatus 700.
The display 703 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch panel, or the like in some embodiments. The display 703 is used for displaying information at the electronic device 700 and for displaying a visual user interface. The components 701 and 703 of the electronic device 700 communicate with each other via a system bus.
In some embodiments of the present invention, when the processor 701 executes an in situ identification procedure of metal substrate corrosion and coating aging in the memory 702, the following steps may be implemented:
Acquiring a damage image of the metal surface, and segmenting the damage image according to a preset target image segmentation model to obtain a plurality of segmentation sub-images;
determining the prior frame size of the prior frame of a plurality of segmentation sub-images based on a preset clustering method, and determining the category information of the prior frame;
constructing an initial matrix corrosion identification model, and training the initial matrix corrosion identification model based on a plurality of segmentation subimages, the prior frame size and the category information to obtain a target matrix corrosion identification model with complete training;
constructing an initial coating aging identification model, and training the initial coating aging identification model based on a plurality of segmentation sub-images, the prior frame size and the category information to obtain a target coating aging identification model;
acquiring an image to be recognized on the surface of a metal, and inputting the image to be recognized into a target image segmentation model to obtain a plurality of sub-images to be recognized;
and identifying a plurality of sub-images to be identified respectively based on the target matrix corrosion identification model and the target coating aging identification model, and correspondingly obtaining a matrix corrosion identification result and a coating aging identification result.
It should be understood that: the processor 701 may perform other functions in addition to the above functions when executing the in-situ identification procedure of metal substrate corrosion and coating aging in the memory 702, which may be specifically referred to the description of the corresponding method embodiment above.
Further, the type of the mentioned electronic device 700 is not specifically limited in the embodiments of the present invention, and the electronic device 700 may be a portable electronic device such as a mobile phone, a tablet computer, a Personal Digital Assistant (PDA), a wearable device, and a laptop computer (laptop). Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices that carry an IOS, android, microsoft, or other operating system. The portable electronic device may also be other portable electronic devices such as laptop computers (laptop) with touch sensitive surfaces (e.g., touch panels) and the like. It should also be understood that in other embodiments of the present invention, the electronic device 700 may not be a portable electronic device, but may be a desktop computer having a touch-sensitive surface (e.g., a touch pad).
Accordingly, the present application also provides a computer readable storage medium, which is used for storing a computer readable program or instruction, and when the program or instruction is executed by a processor, the program or instruction can implement the steps or functions of the in-situ identification method for metal substrate corrosion and coating aging provided by the foregoing method embodiments.
Those skilled in the art will appreciate that all or part of the processes of the methods of the above embodiments may be implemented by instructing relevant hardware (such as a processor, a controller, etc.) by a computer program, and the computer program may be stored in a computer readable storage medium. The computer readable storage medium is a magnetic disk, an optical disk, a read-only memory or a random access memory.
The method and the device for identifying the corrosion of the metal substrate and the aging of the coating in situ provided by the invention are described in detail, the principle and the implementation mode of the invention are explained by applying specific examples, and the description of the examples is only used for helping to understand the method and the core idea of the invention; meanwhile, for those skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (10)
1. An in-situ identification method for corrosion and coating aging of a metal substrate, comprising:
acquiring a damage image of a metal surface, and segmenting the damage image according to a preset target image segmentation model to obtain a plurality of segmentation sub-images;
Determining prior frame sizes of prior frames of the plurality of segmented sub-images based on a preset clustering method, and determining category information of the prior frames;
constructing an initial matrix corrosion recognition model, and training the initial matrix corrosion recognition model based on the plurality of segmentation sub-images, the prior frame size and the category information to obtain a target matrix corrosion recognition model with complete training;
constructing an initial coating aging recognition model, and training the initial coating aging recognition model based on the plurality of segmentation sub-images, the prior frame size and the class information to obtain a target coating aging recognition model;
acquiring an image to be identified on the metal surface, and inputting the image to be identified into the target image segmentation model to obtain a plurality of sub-images to be identified;
and identifying the plurality of sub-images to be identified respectively based on the target matrix corrosion identification model and the target coating aging identification model, and correspondingly obtaining a matrix corrosion identification result and a coating aging identification result.
2. The method of claim 1, wherein after the step of segmenting the damage image according to a preset target image segmentation model to obtain a plurality of segmentation subimages, the method further comprises:
Acquiring a plurality of real sub-images of the damage image, and determining the intersection ratio of the plurality of real sub-images and the plurality of segmentation sub-images;
and determining the segmentation accuracy of the target image segmentation model according to the intersection ratio.
3. The in-situ identification method for metal substrate corrosion and coating aging according to claim 1, characterized in that the target image segmentation model comprises a coding unit and a decoding unit, wherein the coding unit comprises a depth convolution module, a multi-feature extraction module and a first convolution layer; the decoding unit comprises a second convolution layer, a first up-sampling layer, an overlapping layer, a third convolution layer and a second up-sampling layer;
the depth convolution module is used for extracting low-dimensional features and high-dimensional features of the damage image;
the multi-feature extraction module is used for receiving the high-dimensional features and generating combined features;
the first convolution layer is used for performing convolution processing on the combined feature to generate a first convolution feature;
the second convolution layer is used for performing convolution processing on the low-dimensional features to generate second convolution features;
the first upsampling layer is used for upsampling the first convolution feature to obtain a first upsampled feature;
The superposition layer is used for superposing the first upsampling feature and the second convolution feature to obtain a superposition feature;
the third convolution layer is used for carrying out convolution processing on the superposition characteristics to generate third convolution characteristics;
the second upsampling layer is configured to upsample the third convolution feature to obtain the multiple segmentation sub-images.
4. The method of claim 3, wherein the multi-feature extraction module comprises a first sub-convolution layer, a first hole convolution layer, a second hole convolution layer, a third hole convolution layer, a pooling layer, and a stitching layer in parallel;
the convolution kernel of the first sub convolution layer is 1 multiplied by 1 and is used for extracting first feature information of the high-dimensional features;
the convolution kernel of the first cavity convolution layer is 3 multiplied by 3, the sampling rate is 6, and the second feature information of the high-dimensional features is extracted;
the convolution kernel of the second cavity convolution layer is 3 x 3, the sampling rate is 12, and the second cavity convolution layer is used for extracting third feature information of the high-dimensional features;
the convolution kernel of the third void convolution layer is 3 x 3, the sampling rate is 18, and the fourth feature information of the high-dimensional features is extracted;
The pooling layer is used for extracting fifth feature information of the high-dimensional features;
the splicing layer is used for splicing the first feature information, the second feature information, the third feature information, the fourth feature information and the fifth feature information to generate the combined feature.
5. The method of claim 1, wherein the substrate corrosion identification comprises a substrate corrosion category and substrate corrosion parameters, the substrate corrosion category comprises pitting and cracking, and the substrate corrosion parameters comprise a number of pitting, a pitting bore diameter, a pitting bore depth, a breach size, and a cracking depth; the coating aging identification result comprises a coating damage type, wherein the coating damage type comprises mildew, chalking, cracking and falling.
6. The method of in-situ identification of corrosion and coating degradation of a metallic substrate of claim 5, further comprising:
constructing the correlation relationship between the erosion hole number, the erosion hole aperture, the erosion hole depth, the split size and the cracking depth and the corrosion damage degree of the matrix;
And determining the corrosion damage degree of the substrate on the metal surface according to the number of the etching holes, the aperture of the etching holes, the depth of the etching holes, the size of the gap, the cracking depth and the correlation relationship.
7. The method of in-situ identification of metal substrate corrosion and coating degradation of claim 5 wherein said coating degradation identification further comprises a coating HSV color space; the in-situ identification method for the corrosion and the coating aging of the metal substrate further comprises the following steps:
determining similarity of the coating HSV color space and a standard HSV color space;
and determining the coating aging degree of the metal surface according to the similarity.
8. The method of claim 1, further comprising, after said obtaining an image of damage to the metal surface:
and constructing a filtering window, and carrying out filtering processing on the damaged image based on the filtering window to obtain a filtering image.
9. The method of claim 1, wherein the model structure of the target substrate corrosion identification model is a YOLOv3 model structure, and the model structure of the target coating aging identification model is an SSD model structure.
10. An in-situ identification apparatus for corrosion and coating degradation of a metallic substrate, comprising:
the image segmentation module is used for acquiring a damage image of the metal surface, segmenting the damage image according to a preset target image segmentation model and acquiring a plurality of segmentation sub-images;
the prior frame determining module is used for determining the prior frame size of the prior frames of the plurality of segmentation sub-images based on a preset clustering method and determining the category information of the prior frames;
the corrosion identification model building module is used for building an initial matrix corrosion identification model, and training the initial matrix corrosion identification model based on the plurality of segmentation sub-images, the prior frame size and the category information to obtain a target matrix corrosion identification model with complete training;
the coating aging identification model building module is used for building an initial coating aging identification model, and training the initial coating aging identification model based on the plurality of segmentation sub-images, the prior frame size and the category information to obtain a target coating aging identification model;
the to-be-identified image acquisition module is used for acquiring an to-be-identified image of the metal surface, inputting the to-be-identified image into the target image segmentation model and acquiring a plurality of to-be-identified sub-images;
And the identification module is used for identifying the plurality of sub-images to be identified respectively based on the target matrix corrosion identification model and the target coating aging identification model, and correspondingly obtaining a matrix corrosion identification result and a coating aging identification result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210661350.7A CN114757945B (en) | 2022-06-13 | 2022-06-13 | In-situ identification method and device for corrosion of metal substrate and aging of coating |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210661350.7A CN114757945B (en) | 2022-06-13 | 2022-06-13 | In-situ identification method and device for corrosion of metal substrate and aging of coating |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114757945A true CN114757945A (en) | 2022-07-15 |
| CN114757945B CN114757945B (en) | 2022-10-04 |
Family
ID=82336867
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210661350.7A Active CN114757945B (en) | 2022-06-13 | 2022-06-13 | In-situ identification method and device for corrosion of metal substrate and aging of coating |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114757945B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113506239A (en) * | 2021-05-21 | 2021-10-15 | 冶金自动化研究设计院 | Strip steel surface defect detection method based on cross-stage local network |
-
2022
- 2022-06-13 CN CN202210661350.7A patent/CN114757945B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113506239A (en) * | 2021-05-21 | 2021-10-15 | 冶金自动化研究设计院 | Strip steel surface defect detection method based on cross-stage local network |
Non-Patent Citations (2)
| Title |
|---|
| ZHENG NIE: "Analysis On Deeplabv3+ Performance For Automatic Steel Defects Detection", 《ARXIV:2004.04822V2 [CS.CV]》 * |
| 甘胜丰: "《机器视觉表面缺陷检测技术及其在钢铁工业中的应用》", 30 June 2017 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114757945B (en) | 2022-10-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111986159B (en) | Electrode defect detection method and device for solar cell and storage medium | |
| CN106920233B (en) | Scratch detection method, apparatus and electronic equipment based on image procossing | |
| CN112800915B (en) | Building change detection method, device, electronic equipment and storage medium | |
| JP6476802B2 (en) | Information processing apparatus and information processing method | |
| CN113436100B (en) | Method, apparatus, device, medium, and article for repairing video | |
| CN113436162B (en) | Method and device for identifying weld defects on surface of hydraulic oil pipeline of underwater robot | |
| CN113012200B (en) | Method and device for positioning moving object, electronic equipment and storage medium | |
| CN109977750A (en) | Seal true and false method of calibration, device and computer readable storage medium | |
| CN115797350A (en) | Bridge disease detection method and device, computer equipment and storage medium | |
| CN110378421B (en) | Coal mine fire identification method based on convolutional neural network | |
| CN116168351A (en) | Inspection method and device for power equipment | |
| CN112906671B (en) | Method and device for identifying false face-examination picture, electronic equipment and storage medium | |
| CN114359383A (en) | Image positioning method, device, equipment and storage medium | |
| CN114757945B (en) | In-situ identification method and device for corrosion of metal substrate and aging of coating | |
| CN117333443A (en) | Defect detection method and device, electronic equipment and storage medium | |
| CN114937027B (en) | Fan blade defect detection method and device, electronic equipment and storage medium | |
| CN112749701A (en) | Method for generating license plate contamination classification model and license plate contamination classification method | |
| CN115049836B (en) | Image segmentation method, device, equipment and storage medium | |
| CN115810132B (en) | Crack orientation identification method, device, equipment and storage medium | |
| CN115601684A (en) | Emergency early warning method and device, electronic equipment and storage medium | |
| CN112446231A (en) | Pedestrian crossing detection method and device, computer equipment and storage medium | |
| CN113537026B (en) | Method, device, equipment and medium for detecting graphic elements in building plan | |
| CN112906478B (en) | Target object identification method, device, equipment and storage medium | |
| CN114140612A (en) | Method, device, equipment and storage medium for detecting hidden danger of power equipment | |
| CN113610856A (en) | Method and device for training image segmentation model and image segmentation |
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 |