CN114240948A - Intelligent segmentation method and system for structural surface damage image - Google Patents
Intelligent segmentation method and system for structural surface damage image Download PDFInfo
- Publication number
- CN114240948A CN114240948A CN202111327498.9A CN202111327498A CN114240948A CN 114240948 A CN114240948 A CN 114240948A CN 202111327498 A CN202111327498 A CN 202111327498A CN 114240948 A CN114240948 A CN 114240948A
- Authority
- CN
- China
- Prior art keywords
- image
- segmentation
- network
- structural
- structural surface
- 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
- 230000011218 segmentation Effects 0.000 title claims abstract description 71
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000003709 image segmentation Methods 0.000 claims abstract description 84
- 238000007781 pre-processing Methods 0.000 claims abstract description 20
- 238000012549 training Methods 0.000 claims description 30
- 238000012360 testing method Methods 0.000 claims description 27
- 238000012795 verification Methods 0.000 claims description 18
- 230000004927 fusion Effects 0.000 claims description 13
- 238000005070 sampling Methods 0.000 claims description 13
- 230000009191 jumping Effects 0.000 claims description 12
- 238000011176 pooling Methods 0.000 claims description 8
- 239000000284 extract Substances 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 abstract description 3
- 230000006870 function Effects 0.000 description 9
- 238000001514 detection method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 230000000007 visual effect Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000011248 coating agent Substances 0.000 description 1
- 238000000576 coating method Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 238000011065 in-situ storage Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000009659 non-destructive testing Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000009991 scouring Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an intelligent segmentation method of a structural surface damage image, which comprises the following steps: acquiring a plurality of structural surface images to be segmented, and preprocessing each structural surface image; and inputting the preprocessed structure surface image into the trained image segmentation network to obtain the damage condition of the structure surface. By the method, damage misjudgment caused by background pixels can be avoided in the structural damage image segmentation processing process, and damage identification accuracy is improved.
Description
Technical Field
The invention belongs to the technical field of nondestructive testing and computer vision, and particularly relates to an intelligent segmentation method and system for a structural surface damage image.
Background
A large number of relatively independent key structural components, such as blades, rotors, heat insulating tiles and the like, exist in heavy equipment such as gas turbines, steam turbines, aircraft engines, nuclear power plants and the like. The structures are often subjected to complex working conditions such as high temperature, high pressure, airflow scouring and the like in the service process, surface damages such as corrosion, cracks, pits, coating peeling and the like are easily generated, and the service safety of the equipment is seriously influenced. The regular inspection and evaluation of critical structures, the advance prediction of potentially dangerous areas and the repair or replacement of them are important requirements for the operation and maintenance of equipment. The conventional structure surface damage assessment mainly adopts manual visual detection, the method seriously depends on the attention state and the working experience of field workers, and has poor stability and consistency and extremely low efficiency. The method for detecting the surface damage of the large-batch structures in different categories by adopting the intelligent image identification method is great trend, is expected to replace manual visual detection, and greatly improves the accuracy and efficiency of damage identification. The effective segmentation of the image features is the basis for realizing accurate identification. The traditional image segmentation method mainly uses the significant difference of characteristics such as gray scale, color, texture and the like to divide different areas, and the common method comprises the following steps: a threshold segmentation method, a boundary segmentation method, a region segmentation method, a watershed segmentation method, and the like. The methods rely on manual feature extraction and description, are difficult to obtain good effect in complex environment; and the method has the advantages of strong limitation, poor robustness, low stability, no universality and no effect of automatic segmentation. With the development of intelligent technology in the field of computer vision, research on analyzing digital images by utilizing a deep convolutional neural network is remarkably advanced. In 2015, researchers provide a classic image segmentation network FCN, so that pixel-level segmentation of an image on a specific class target object is realized, and the deeper requirements of image analysis are met. However, the damaged pixels segmented by the FCN method easily overflow the structural size range, and the background pixels outside the structural range are considered as damaged features, thereby causing erroneous judgment. In the actual digital image acquisition, many structures can only be shot in situ due to the difficulty in disassembly, the background is more complex, and the possibility of misjudgment is higher. Measures are necessary to eliminate the misjudgment of the background pixels outside the structural range, and the accuracy and consistency of the structural damage image segmentation are improved.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide an intelligent segmentation method and system for a structural surface damage image, which can solve the problem that a damage segmentation result overflows a structural range by integrating mask operation into an image segmentation network, and improve the accuracy and consistency of the segmentation result.
In order to achieve the purpose, the invention provides the following technical scheme:
an intelligent segmentation method for a structural surface damage image comprises the following steps:
s100: acquiring a plurality of structural surface images to be segmented, and preprocessing each structural surface image;
s200: and inputting each preprocessed structural surface image into the trained image segmentation network to obtain the damage condition of the structural surface.
Preferably, the image segmentation network comprises a first segmentation module and a second segmentation module,
the first segmentation module is a feature extraction network, extracts the features in each preprocessed structure surface image through convolution and pooling, and performs first segmentation on the structure surface image through up-sampling and jumping fusion operation to obtain a structure area;
the second segmentation module takes the structural region obtained by the first segmentation module as a label, removes background information through a covering mask, and performs secondary segmentation on the structural region through an up-sampling and jumping fusion operation to obtain a damaged region of the structure surface.
Preferably, the image segmentation network is trained by:
s1000: acquiring a plurality of structural surface images with damage, preprocessing the structural surface images, and forming a training set, a verification set and a test set of the preprocessed structural surface images;
s2000: marking each structural surface image in the training set, inputting an image segmentation network to train the network, inputting each structural surface image in the verification set into the trained image segmentation network to verify the network, and finishing the training of the image segmentation network when the score of the verification set is not increased any more;
s3000: and inputting each image in the test set into the trained image segmentation network to test the network, and if the test is passed, obtaining the final image segmentation network.
Preferably, in step S1000, a plurality of structural surface images are preprocessed: and carrying out scaling processing on each image by adopting a bilinear interpolation method.
The invention also provides an intelligent segmentation system for the structural surface damage image, which comprises the following steps:
the image input unit is used for inputting a structure surface image to be segmented;
the image preprocessing unit is used for preprocessing the surface image of the structure to be segmented;
and the image segmentation unit is used for carrying out image segmentation on the preprocessed structure surface image through an image segmentation network so as to obtain the damage condition of the structure surface.
Preferably, the image segmentation unit includes a first segmentation module and a second segmentation module,
the first segmentation module is a feature extraction network, extracts the features in each preprocessed structure surface image through convolution and pooling, and performs first segmentation on the structure surface image through up-sampling and jumping fusion operation to obtain a structure area;
the second segmentation module takes the structural region obtained by the first segmentation module as a label, removes background information in the feature through a covering mask, and performs secondary segmentation on the structural region through an up-sampling and jumping fusion operation to obtain a damaged region of the structure surface.
Preferably, the image segmentation unit further includes an image segmentation network training module, configured to train the image segmentation network.
Preferably, the image segmentation network training module includes:
the image input and preprocessing submodule is used for acquiring a plurality of structural surface images with damage, preprocessing the structural surface images, and forming a training set, a verification set and a test set of the preprocessed structural surface images;
the image training submodule is used for marking each structural surface image in the training set, inputting an image segmentation network to train the network, inputting each structural surface image in the verification set into the trained image segmentation network to verify the network, and finishing the training of the image segmentation network when the score of the verification set is not increased any more;
and the image testing sub-module is used for inputting each image in the test set into the trained image segmentation network to test the network, and if the test is passed, the final image segmentation network is obtained.
The present invention also provides a computer apparatus comprising:
a memory and a processor, wherein,
the memory has stored thereon an executable program operable on the processor,
the processor executes the executable program to realize the structure surface damage image segmentation method.
The invention also provides a computer readable storage medium, which stores a computer program, and the computer program can realize the structure surface damage image segmentation method when being executed by a processor.
Compared with the prior art, the invention has the following beneficial effects:
compared with manual visual detection, the visual detection device can assist manual visual detection of structural surface damage, ensure consistency and accuracy and improve production efficiency;
secondly, compared with traditional image processing methods such as threshold-based segmentation and the like, the method provided by the invention solves the problem of high difficulty in manually extracting the features, and has strong universality and good robustness.
Thirdly, compared with a classic intelligent image segmentation network represented by FCN, the method is improved on the basis of the intelligent image segmentation network, a two-step progressive strategy with limited range is provided, the problem that damage segmentation results overflow the structural range is solved, and the accuracy of the segmentation results is improved.
Drawings
FIG. 1 is a flowchart of a method for intelligently segmenting a damage image of a surface of a structure according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of an image segmentation network according to another embodiment of the present invention;
FIG. 3 is a flow chart of network training provided by another embodiment of the present invention;
FIG. 4 is a diagram illustrating the segmented image of the blade surface according to another embodiment of the present invention;
fig. 5(a) is a schematic diagram of a blade surface damage after being segmented by an FCN network according to another embodiment of the present invention, and fig. 5(b) is a schematic diagram of a blade surface damage after being segmented by an image segmentation network according to another embodiment of the present invention.
Detailed Description
Specific embodiments of the present invention will be described in detail below with reference to fig. 1 to 5 (b). While specific embodiments of the invention are shown in the drawings, it should be understood that the invention may be embodied in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.
It should be noted that certain terms are used throughout the description and claims to refer to particular components. As one skilled in the art will appreciate, various names may be used to refer to a component. This specification and claims do not intend to distinguish between components that differ in name but not function. In the following description and in the claims, the terms "include" and "comprise" are used in an open-ended fashion, and thus should be interpreted to mean "include, but not limited to. The description which follows is a preferred embodiment of the invention, but is made for the purpose of illustrating the general principles of the invention and not for the purpose of limiting the scope of the invention. The scope of the present invention is defined by the appended claims.
For the purpose of facilitating understanding of the embodiments of the present invention, the following description will be made by taking specific embodiments as examples with reference to the accompanying drawings, and the drawings are not to be construed as limiting the embodiments of the present invention.
In one embodiment, as shown in fig. 1, taking a surface damage image of a heavy-duty gas turbine blade as an example, the invention provides an intelligent segmentation method for a structural surface damage image, which includes the following steps:
s100: acquiring a plurality of blade surface images to be segmented, and preprocessing each blade surface image;
s200: and inputting the preprocessed surface images of each blade into the trained image segmentation network to obtain the damage condition of the blade surface.
In this embodiment, as shown in fig. 2, the image segmentation network comprises a first segmentation module and a second segmentation module,
the first segmentation module is a feature extraction network, extracts the features in each preprocessed blade surface image through convolution and pooling, and performs first segmentation on the blade surface image through up-sampling and jumping fusion operations to obtain a blade area;
the second segmentation module takes the blade region obtained by the first segmentation module as a label, removes background information in the feature through a covering mask, and performs secondary segmentation on the blade region through an up-sampling and jumping fusion operation to obtain a damaged region on the surface of the blade.
In another embodiment, the image segmentation network is trained by:
s1000: acquiring a plurality of damaged blade surface images, preprocessing the images, and forming a training set, a verification set and a test set on the preprocessed blade surface images;
in this step, after acquiring the surface images of the plurality of blades, the images are preferably preprocessed, specifically, in the following manner:
preprocessing the surface images of the multiple leaves: and carrying out scaling processing on each image by adopting a bilinear interpolation method.
S2000: marking the surface images of all the leaves in the training set, inputting an image segmentation network to train the network, inputting the surface images of all the leaves in the verification set into the trained image segmentation network to verify the network, and finishing the training of the image segmentation network when the scores of the verification set are not increased any more;
in this step, before training the image segmentation network, it is necessary to mark each leaf surface image in the training set with a different label, as shown in table 1:
TABLE 1
In table 1, a label (i) indicates a leaf and a background area, where the leaf area is set to 1 and the background area is set to 0; and a label II represents a damaged area and a non-damaged area on the blade, wherein the damaged area is set to be 1, and the non-damaged area is set to be 0.
Inputting the marked blade surface image into an image segmentation network, wherein the image segmentation network extracts the characteristics of a blade region and a blade damage region together after a series of convolution and pooling operations are performed on the marked image, and then performs an up-sampling operation, jumps and fuses the characteristic maps of all pooling layers and outputs the blade region as shown in fig. 3; then, as shown in fig. 4, the leaf label is used as a mask, the mask is covered on the feature map, and the background region other than the leaf region on the feature map is set to 0 by the following formula, thereby obtaining a background-removed feature map.
Where, (x, y) denotes a point on the feature map, Φ ═ x, y denotes a feature value, a denotes a leaf region, and B denotes a background region.
Further, the background removing feature map is also subjected to upsampling and jump fusion operation, so that a blade damage area is output.
After obtaining the blade area and the blade damage area, the predicted values of the blade area and the blade damage area need to be compared with their labels (true values) through a loss function, where the loss function is shown as follows:
L＝La+λLb
wherein L represents the overall loss function, LaLoss function, L, obtained by representing predicted and true values of the bladebAnd expressing a loss function obtained by the predicted value and the true value of the damage, and expressing a parameter for adjusting the proportion of the two loss functions by lambda.
And calculating a loss function by combining two output results of the blade area and the blade damage area, and updating learnable parameters in the model according to a back propagation algorithm by taking the minimum loss function L as a target so as to obtain an optimal image segmentation network.
S3000: and inputting each image in the test set into the trained image segmentation network to test the network, and if the test is passed, obtaining the final image segmentation network.
Fig. 5(a) and 5(b) are schematic diagrams illustrating the effect of the original blade surface image after being segmented by the image segmentation network and the FCN network according to the present invention, wherein fig. 5(a) is the segmentation result of the FCN network model, and fig. 5(b) is the segmentation result of the image segmentation network according to the present invention. It can be seen that in fig. 5(a), the segmentation results clearly overflow the range of the leaf region, whereas in fig. 5(b), the segmentation results are only performed within the range of the leaf region, so that the following conclusion is not difficult to draw: the image segmentation network provided by the invention can avoid the segmentation result from overflowing the leaf range, thereby improving the accuracy of image segmentation.
It should be noted that the above-mentioned embodiments are only preferred embodiments of the present invention, and are not intended to limit the present invention. The image segmentation network and the method for segmenting the image by using the same are suitable for the blade surface damage image and other structures with surface damage.
In another embodiment, the present invention further provides an intelligent segmentation system for a damage image of a surface of a structure, comprising:
the image input unit is used for inputting a blade surface image to be segmented;
the image preprocessing unit is used for preprocessing the surface image of the blade to be segmented;
and the image segmentation unit is used for carrying out image segmentation on the preprocessed blade surface image through an image segmentation network so as to obtain the damage condition of the blade surface.
In this embodiment, the image segmentation unit includes a first segmentation module and a second segmentation module,
the first segmentation module is a feature extraction network, extracts the features in each preprocessed blade surface image through convolution and pooling, and performs first segmentation on the blade surface image through up-sampling and jumping fusion operations to obtain a blade area;
the second segmentation module takes the blade region obtained by the first segmentation module as a label, removes background information in the feature through a covering mask, and performs secondary segmentation on the blade region through an up-sampling and jumping fusion operation to obtain a damaged region on the surface of the blade.
In another embodiment, the image segmentation unit further includes an image segmentation network training module, configured to train the image segmentation network.
In another embodiment, the image segmentation network training module includes:
the image input and preprocessing submodule is used for acquiring a plurality of damaged blade surface images, preprocessing the images, and forming a training set, a verification set and a test set on the preprocessed blade surface images;
the image training submodule is used for marking the surface images of all the blades in the training set, inputting an image segmentation network to train the network, inputting the surface images of all the blades in the verification set into the trained image segmentation network to verify the network, and finishing the training of the image segmentation network when the scores of the verification set are not increased any more;
and the image testing sub-module is used for inputting each image in the test set into the trained image segmentation network to test the network, and if the test is passed, the final image segmentation network is obtained.
In another embodiment, the present invention also provides a computer apparatus comprising:
a memory and a processor, wherein,
the memory has stored thereon an executable program operable on the processor,
the processor executes the executable program to implement the segmentation method of the blade surface damage image as in the above embodiments.
In another embodiment, a computer-readable storage medium stores a computer program which, when executed by a processor, implements the method for segmenting a blade surface damage image as in the above embodiments.
In this embodiment, it should be understood by those skilled in the art that the method described in the above embodiment can be implemented by instructing relevant hardware through a computer program, and when the computer program is executed, all the steps of the above method can be included. Any reference to memory, storage, database or other medium used in various embodiments provided by this embodiment may include any type of non-volatile memory known in the art, including read only memory ROM, programmable memory PROM, electrically programmable memory EPROM, and any type of volatile memory, including static RAM, dynamic RAM, synchronous DRAM, and the like.
Claims (10)
1. An intelligent segmentation method and system for a structural surface damage image comprises the following steps:
s100: acquiring a plurality of structural surface images to be segmented, and preprocessing each structural surface image;
s200: and inputting each preprocessed structural surface image into the trained image segmentation network to obtain the damage condition of the structural surface.
2. The method of claim 1, wherein preferably the image segmentation network comprises a first segmentation module and a second segmentation module,
the first segmentation module is a feature extraction network, extracts the features in each preprocessed structure surface image through convolution and pooling, and performs first segmentation on the structure surface image through up-sampling and jumping fusion operation to obtain a structure area;
the second segmentation module takes the structural region obtained by the first segmentation module as a label, removes background information in the feature through a covering mask, and performs secondary segmentation on the structural region through an up-sampling and jumping fusion operation to obtain a damaged region of the structure surface.
3. The method of claim 1, wherein the image segmentation network is trained by:
s1000: acquiring a plurality of structural surface images with damage, preprocessing the structural surface images, and forming a training set, a verification set and a test set of the preprocessed structural surface images;
s2000: marking each structural surface image in the training set, inputting an image segmentation network to train the network, inputting each structural surface image in the verification set into the trained image segmentation network to verify the network, and finishing the training of the image segmentation network when the score of the verification set is not increased any more;
s3000: and inputting each image in the test set into the trained image segmentation network to test the network, and if the test is passed, obtaining the final image segmentation network.
4. The method according to claim 3, wherein in step S1000, a plurality of structural surface images are preprocessed by: and (5) zooming the image by adopting a bilinear interpolation method.
5. An intelligent segmentation system for images of damage to a surface of a structure, comprising:
the image input unit is used for inputting a structure surface image to be segmented;
the image preprocessing unit is used for preprocessing the surface image of the structure to be segmented;
and the image segmentation unit is used for carrying out image segmentation on the preprocessed structure surface image through an image segmentation network so as to obtain the damage condition of the structure surface.
6. The system of claim 5, wherein the image segmentation unit comprises a first segmentation module and a second segmentation module,
the first segmentation module is a feature extraction network, extracts the features in each preprocessed structure surface image through convolution and pooling, and performs first segmentation on the structure surface image through up-sampling and jumping fusion operation to obtain a structure area;
the second segmentation module takes the structural region obtained by the first segmentation module as a label, removes background information in the feature through a covering mask, and performs secondary segmentation on the structural region through an up-sampling and jumping fusion operation to obtain a damaged region of the structure surface.
7. The system of claim 5, wherein the image segmentation unit further comprises an image segmentation network training module to train an image segmentation network.
8. The system of claim 5, wherein the image segmentation network training module comprises:
the image input and preprocessing submodule is used for acquiring a plurality of structural surface images with damage, preprocessing the structural surface images, and forming a training set, a verification set and a test set of the preprocessed structural surface images;
the image training submodule is used for marking each structural surface image in the training set, inputting an image segmentation network to train the network, inputting each structural surface image in the verification set into the trained image segmentation network to verify the network, and finishing the training of the image segmentation network when the score of the verification set is not increased any more;
and the image testing sub-module is used for inputting each image in the test set into the trained image segmentation network to test the network, and if the test is passed, the final image segmentation network is obtained.
9. A computer device, comprising:
a memory and a processor, wherein,
the memory has stored thereon an executable program operable on the processor,
the processor executes the executable program to implement the method of any one of claims 1-4.
10. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, is adapted to carry out the method according to any one of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111327498.9A CN114240948B (en) | 2021-11-10 | 2021-11-10 | Intelligent segmentation method and system for structural surface damage image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111327498.9A CN114240948B (en) | 2021-11-10 | 2021-11-10 | Intelligent segmentation method and system for structural surface damage image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114240948A true CN114240948A (en) | 2022-03-25 |
| CN114240948B CN114240948B (en) | 2024-03-05 |
Family
ID=80748925
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111327498.9A Active CN114240948B (en) | 2021-11-10 | 2021-11-10 | Intelligent segmentation method and system for structural surface damage image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114240948B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111178392A (en) * | 2019-12-10 | 2020-05-19 | 中国民航大学 | Aero-engine hole-exploring image damage segmentation method based on deep neural network |
| WO2020119103A1 (en) * | 2018-12-13 | 2020-06-18 | 程琳 | Aero-engine hole detection image damage intelligent identification method based on deep learning |
| CN111681233A (en) * | 2020-06-11 | 2020-09-18 | 北京小白世纪网络科技有限公司 | US-CT image segmentation method, system and equipment based on deep neural network |
| CN111696075A (en) * | 2020-04-30 | 2020-09-22 | 航天图景（北京）科技有限公司 | Intelligent fan blade defect detection method based on double-spectrum image |
| CN113096085A (en) * | 2021-04-01 | 2021-07-09 | 武汉理工大学 | Container surface damage detection method based on two-stage convolutional neural network |
-
2021
- 2021-11-10 CN CN202111327498.9A patent/CN114240948B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020119103A1 (en) * | 2018-12-13 | 2020-06-18 | 程琳 | Aero-engine hole detection image damage intelligent identification method based on deep learning |
| CN111178392A (en) * | 2019-12-10 | 2020-05-19 | 中国民航大学 | Aero-engine hole-exploring image damage segmentation method based on deep neural network |
| CN111696075A (en) * | 2020-04-30 | 2020-09-22 | 航天图景（北京）科技有限公司 | Intelligent fan blade defect detection method based on double-spectrum image |
| CN111681233A (en) * | 2020-06-11 | 2020-09-18 | 北京小白世纪网络科技有限公司 | US-CT image segmentation method, system and equipment based on deep neural network |
| CN113096085A (en) * | 2021-04-01 | 2021-07-09 | 武汉理工大学 | Container surface damage detection method based on two-stage convolutional neural network |
Non-Patent Citations (1)
| Title |
|---|
| 张维亮;李楠;李昂;石宏;: "基于概率神经网络的航空发动机叶片损伤图像分割", 沈阳航空航天大学学报, no. 02 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114240948B (en) | 2024-03-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111047554B (en) | Composite insulator overheating defect detection method based on instance segmentation | |
| CN109493320B (en) | Remote sensing image road extraction method and system based on deep learning, storage medium and electronic equipment | |
| CN111563893B (en) | Grading ring defect detection method, device, medium and equipment based on aerial image | |
| Schmugge et al. | Crack segmentation by leveraging multiple frames of varying illumination | |
| CN106056619A (en) | Unmanned aerial vehicle vision wire patrol method based on gradient constraint Radon transform | |
| CN111798412A (en) | Intelligent diagnosis method and system for substation equipment defects based on infrared images | |
| CN111178392B (en) | Aero-engine hole detection image damage segmentation method based on deep neural network | |
| Sun et al. | Research on fatigue crack growth detection of M (T) specimen based on image processing technology | |
| CN114663672B (en) | Method and system for detecting corrosion of steel member of power transmission line tower | |
| CN112906689A (en) | Image detection method based on defect detection and segmentation depth convolution neural network | |
| CN112884715A (en) | Composite insulator grading ring inclination fault detection method based on deep learning | |
| CN115995056A (en) | Automatic bridge disease identification method based on deep learning | |
| CN115239646A (en) | Defect detection method and device for power transmission line, electronic equipment and storage medium | |
| CN108537792B (en) | Power defect image identification method based on convolutional neural network | |
| CN113962929A (en) | Photovoltaic cell assembly defect detection method and system and photovoltaic cell assembly production line | |
| CN105787955A (en) | Sparse segmentation method and device of strip steel defect | |
| CN114240948A (en) | Intelligent segmentation method and system for structural surface damage image | |
| CN116342542A (en) | Lightweight neural network-based steel product surface defect detection method | |
| CN115953387A (en) | Radiographic image weld defect detection method based on deep learning | |
| CN116167999A (en) | Distribution line zero-value insulator infrared thermal image detection method based on image matching | |
| CN115689995A (en) | Circular hole edge sub-pixel level extraction method under complex illumination background condition | |
| CN112036410B (en) | Electric power equipment infrared image segmentation method based on deep learning and image gradient | |
| CN115330705A (en) | Skin paint surface defect detection method based on adaptive weighting template NCC | |
| CN110223346B (en) | Shape prior-based multi-insulator positioning method | |
| Chen et al. | Surface Defect Detection of Cable Based on Threshold Image Difference |
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 |