CN111861906A - Pavement crack image virtual augmentation model establishment and image virtual augmentation method - Google Patents
Pavement crack image virtual augmentation model establishment and image virtual augmentation method Download PDFInfo
- Publication number
- CN111861906A CN111861906A CN202010574126.5A CN202010574126A CN111861906A CN 111861906 A CN111861906 A CN 111861906A CN 202010574126 A CN202010574126 A CN 202010574126A CN 111861906 A CN111861906 A CN 111861906A
- Authority
- CN
- China
- Prior art keywords
- pavement crack
- model
- crack image
- network
- image
- 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
- 230000003416 augmentation Effects 0.000 title claims abstract description 33
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000012549 training Methods 0.000 claims abstract description 23
- 230000006870 function Effects 0.000 claims abstract description 20
- 230000004913 activation Effects 0.000 claims description 6
- 230000003190 augmentative effect Effects 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 2
- 238000012545 processing Methods 0.000 abstract description 3
- 238000003709 image segmentation Methods 0.000 abstract 1
- 230000000694 effects Effects 0.000 description 18
- 238000001514 detection method Methods 0.000 description 12
- 238000010586 diagram Methods 0.000 description 9
- 230000008859 change Effects 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000004042 decolorization Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 230000017105 transposition Effects 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 238000012935 Averaging Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 230000009916 joint effect Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 238000009827 uniform distribution Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- 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
- G06T5/00—Image enhancement or restoration
- G06T5/90—Dynamic range modification of images or parts thereof
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to the field of pavement crack image processing, and discloses a pavement crack image virtual augmentation model building and image virtual augmentation method. The model establishing method comprises the following steps: step 1: acquiring a pavement crack image, and sequentially performing data quality improvement and image segmentation on the pavement crack image to obtain a real pavement crack image set; step 2: establishing a DCGAN generation confrontation network model, wherein the DCGAN generation confrontation network model comprises a generator network and a discriminator network, and penalty terms are set behind loss functions of the generator network and the discriminator network; and step 3: and (3) acquiring random noise, inputting the random noise and the real pavement crack image set obtained in the step (1) into the DCGAN generation countermeasure network model obtained in the step (2) for training, wherein the trained model is the pavement crack image virtual augmentation model. The method effectively solves the problem of insufficient crack image data sets, and well realizes the increase of the number and the diversity of the crack image data sets.
Description
Technical Field
The invention belongs to the field of pavement crack image processing, and particularly relates to a pavement crack image virtual augmentation model building and image virtual augmentation method.
Background
With the combination of the computer vision algorithm and the road image processing field, the artificial intelligence auxiliary pavement disease detection and maintenance become possible. In practical application, a crack detection model with good performance usually needs a large-scale and strong-diversity data set, but the acquisition of a crack image is a time-consuming and labor-consuming project, and the data requirement of the current deep learning network model is difficult to meet. Although the image data set is usually augmented by conventional methods such as turning, translation and the like at the present stage, the number of data set samples augmented by the method is limited and the diversity is poor, and if the data set is used for training models such as pavement crack detection, classification and segmentation, the trained network model lacks the generalization performance and is easy to generate the over-fitting phenomenon. Therefore, how to improve the phenomena of poor generalization capability of network models and model overfitting by improving the image augmentation method becomes a problem to be solved urgently.
Disclosure of Invention
The invention aims to provide a method for establishing a virtual augmentation model of a pavement crack image and a virtual augmentation method of the image, which are used for solving the problems that the augmentation technology in the prior art cannot augment the pavement crack in a large amount and high efficiency and the like.
In order to realize the task, the invention adopts the following technical scheme:
a method for establishing a virtual augmentation model of a pavement crack image comprises the following steps:
step 1: collecting a pavement crack image, and preprocessing the pavement crack image to obtain a real pavement crack image set, wherein the pavement crack is a transverse seam or a longitudinal seam;
step 2: establishing a DCGAN generation confrontation network model, wherein the DCGAN generation confrontation network model comprises a generator network and a discriminator network, and penalty terms are set behind loss functions of the generator network and the discriminator network;
and step 3: and (3) acquiring random noise, inputting the random noise and the real pavement crack image set acquired in the step (1) into the DCGAN generation countermeasure network model acquired in the step (2) for training, wherein the generator network model after training is the pavement crack image virtual augmentation model.
Furthermore, the generator network has 5 layers in total, the first layer is a full connection layer, the second layer to the fifth layer are convolution layers, and the convolution layers all adopt Relu as an activation function.
Furthermore, the discriminator network has 5 layers, the first to fourth layers are convolution layers, the fifth layer is a full connection layer, and the convolution layers all adopt leakyrelu as an activation function.
Further, the penalty function after adding the penalty term is
And is
Wherein X represents the input of the generator network and the discriminator network, y represents the output of the generator network and the discriminator network, w is a weight vector, alpha represents the value range of the regularization coefficient from 0 to 1, | | w | |1Is the sum of the absolute values of the weights w in the model.
Further, the preprocessing in step 1 includes operations of color removal, noise removal, contrast enhancement, brightness enhancement and division into uniform sizes, which are performed in sequence.
An image virtual augmentation method comprising the steps of:
and acquiring random noise, and inputting the random noise into the pavement crack image virtual augmentation model acquired by any pavement crack image virtual augmentation model establishing method to acquire a pavement crack image augmentation data set.
Compared with the prior art, the invention has the following technical characteristics:
(1) the method uses the deep convolution generation type countermeasure network to generate the pavement crack image for the first time. The strategy of using the L1 regularization constraint is selected to reduce the test error in order to reduce the learning of the network to noise, enhance the generalization capability of the model, make the model smoother and prevent the over-fitting phenomenon.
(2) According to the method, the pavement crack image with more vividness and high quality is obtained by optimizing each hyper-parameter in the countermeasure network generated by adopting the deep convolution.
(3) The method provided by the invention effectively solves the problem of insufficient crack image data sets, and well realizes the expansion of the number and diversity of the crack image data sets.
Drawings
FIG. 1 is a general flow chart of the present invention;
FIG. 2 is a schematic diagram of a generative countermeasure network architecture;
FIG. 3 is a diagram of a generator network structure in a deep convolution generation countermeasure network;
FIG. 4 is a diagram of a discriminator network structure in a deep convolution generated countermeasure network;
FIG. 5 is a diagram showing the effect of generating transverse cracks in the crack growth model; wherein (a) is a 100epoch generated effect graph; (b) generating an effect graph for 200 epochs; (c) generating an effect map for 300 epoch; (d) generating an effect map for 400 epoch; (e) generating an effect map for 500 epochs; (f) generating an effect graph for 600 epoch;
FIG. 6 is a graph showing the effect of longitudinal seam formation in the crack growth model; wherein (a) is a 100epoch generated effect graph; (b) generating an effect graph for 200 epochs; (c) generating an effect map for 300 epoch; (d) generating an effect map for 400 epoch; (e) generating an effect map for 500 epochs; (f) generating an effect graph for 600 epoch;
FIG. 7 is a graph of the change in crack detection accuracy before and after data set expansion.
The details of the present invention are explained in further detail below with reference to the drawings and examples.
Detailed Description
The following embodiments of the present invention are provided, and it should be noted that the present invention is not limited to the following embodiments, and all equivalent changes based on the technical solutions of the present invention are within the protection scope of the present invention.
Example 1
The embodiment discloses a method for establishing a virtual augmentation model of a pavement crack image, which comprises the following steps:
step 1: collecting a pavement crack image, and preprocessing the pavement crack image to obtain a real pavement crack image set, wherein the pavement crack is a transverse seam or a longitudinal seam;
step 2: establishing a DCGAN generation confrontation network model, wherein the DCGAN generation confrontation network model comprises a generator network and a discriminator network, and penalty terms are set behind loss functions of the generator network and the discriminator network;
and step 3: and (3) acquiring random noise, inputting the random noise and the real pavement crack image set acquired in the step (1) into the DCGAN generation countermeasure network model acquired in the step (2) for training, wherein the generator network model after training is the pavement crack image virtual augmentation model.
The generator network is used for inputting random noise and outputting to generate a pavement crack image set, and the discriminator is used for inputting the generated pavement crack image set and the real pavement crack image set and outputting a real probability value.
Specifically, the random noise is generated by using TensorFlow.
Specifically, the preprocessing in step 1 includes operations of color removal, noise removal, contrast enhancement, brightness enhancement and division into uniform sizes, which are performed in sequence, and the quality of the preprocessed image data is improved.
Specifically, the real probability value in step 3 can represent the similarity between the picture generated by the generator and the real picture, and if the probability value is greater than 0.5, the input of the real probability value can be determined to be from the real data set; if the probability value is less than 0.5, the input is the newly generated data of the generation model;
specifically, the loss function of each training batch of the model is shown in formula (1),
is the true probability value, y, output by the jth training batch of the discriminator networkjThe model is a prediction probability value of the model, N is the total number of training batches, and the loss function value of the model is obtained by summing the loss function values of each batch and then averaging:
specifically, a gradient descent method is employed herein to update the values of the weight w and the offset b in each layer.
Specifically, the gradient is obtained through a back propagation algorithm, and in addition, in order to improve the training speed, the training data is divided into small batches (mini-batches). In each small batch, a small portion of the training samples is randomly selected at each step for gradient calculation. This has the advantage of increasing the speed of training.
Each cycle is defined as the process of a single full training of the training data, requiring multiple such cycles for the entire training process. At the end of each cycle, the penalty function used on the validation set is evaluated, and the network that minimizes the validation set penalty function is selected as the best choice. And training the generator according to the size of the loss function to guide the training direction of the model.
In the deep convolutional neural network, sample data is the basis of network training, and regularization constraints are guidelines for ensuring the network learning direction. In order to reduce the learning of noise by the network, reduce the test error, enhance the generalization capability of the model, make the model smoother and prevent the over-fitting phenomenon, the invention selects the strategy of using the L1 regularization constraint.
Specifically, the loss function after adding the penalty term is
And is
Wherein X represents the input of the generator network and the discriminator network, y represents the output of the generator network and the discriminator network, w is a weight vector, alpha represents the value range of the regularization coefficient from 0 to 1, | | w | | 1Is the sum of the absolute values of the weights w in the model.
Graduating the above equation yields:
when w is positive, w becomes smaller after updating; when w is negative, updated w becomes larger. The distance is as close to 0 as possible, so that the complexity of the network is reduced, and the over-fitting phenomenon is prevented. When w is 0, | w | is not derivable, sign (0) is made 0, and the case where w is 0 is included in the regularization constraint.
The L1 regularization utilizes the sparse characteristic to carry out the feature selection, and selects the meaningful features from the feature subset, and the characteristic greatly simplifies the deep learning feature extraction problem, so that the generating effect of the generator is better.
Specifically, the generator network has 5 layers in total, wherein the convolutional layer is 4 layers, the fully-connected layer is 1 layer, and the specific operation is to use a random variable which is 100-dimensional and subject to (0,1) uniform distribution as the input of the generator. Firstly, the image is processed by a full connection layer to obtain a 4 x 1024 image, and then processed by 4 convolution kernels and a 5 x 5 transposition convolution layer to continuously enlarge the image, and finally, a crack image output by a generator is obtained. In the whole generator structure, the Relu activation function is used by other layers in the network except the first layer.
Specifically, the input of the discriminator is the generated crack image and the real crack image, and the output is a probability value for judging the authenticity, namely the value can judge the authenticity of the input image. The discriminator network used herein has a total of 5 layers, of which the convolutional layer is 4 layers and the fully connected layer is 1 layer. 64 × 64 × 3 images at the input of the discriminator. Firstly, the image is continuously downsampled for 4 times, the step length is 2, the size of a convolution kernel is 5 multiplied by 5, so that the image can be gradually reduced, and finally, a probability value representing the authenticity of the input image is obtained through a full connection layer. In the whole discriminator structure, the leakyrlelu function is adopted by other layers except the last layer without using the activation function.
Preferably, the optimizers of the generator and the arbiter are set to Adam optimizers, and the learning rates are set to 0.0002, and the convolution kernel sizes are 5 × 5.
Table 1 shows the network parameter settings for the deep convolution-generated countermeasure network constructed by the present invention;
table 1 DCGAN network parameter settings
FIG. 2 is a schematic diagram of a structure of a generative countermeasure network, which is mainly composed of two parts, including a generator and a discriminator
Fig. 3 is a diagram showing a structure of a generator network. The network structure is deconvolution (Decpmovavolvulsm), which is generally called Transposed Convolution (Transposed Convolution), that is, feature map size is gradually increased by means of learning. The input of the generated model in the training process is generally random noise Z, and new data G (Z) which is consistent with the real data distribution as much as possible is output directionally or non-directionally through a plurality of layers of transposition convolution calculation. Whether a generating network can produce high quality results depends on how the structure and parameters of the generating network are optimized.
Fig. 4 is a diagram showing the structure of the arbiter network. The network structure is similar to the traditional cyclic convolution neural network, and the features of the input image are extracted mainly through a plurality of convolutions, so that the size of the feature map of the hidden layer is generally reduced gradually. By inputting the real data set X and the new data G (z) generated by the generation model at the same time, obtaining an output probability value to judge whether the input of the input comes from the real data set, and if the probability value is more than 0.5, determining that the input comes from the real data set; if the probability value is less than 0.5, the input is the newly generated data of the generation model.
The method selects a Faster R-CNN detection model to evaluate the quality of the generated crack image and the reliability of the enhanced crack image data set.
Example 2
The embodiment discloses an image virtual augmentation method, which comprises the following steps:
and acquiring random noise, and inputting the random noise into the pavement crack image virtual augmentation model acquired by any pavement crack image virtual augmentation model establishing method to acquire a pavement crack image augmentation data set.
FIG. 5 is a cross-road joint effect diagram generated by the final optimized augmentation model. It can be seen from fig. 5 that the road surface transverse seam effect graph generated by the anti-network model generated by deep convolution becomes more and more vivid as the iteration number increases.
FIG. 6 is a road surface longitudinal seam effect diagram generated by the finally optimized augmentation model. From fig. 6, it can be seen that the pavement longitudinal seam effect graph generated by the anti-network model generated by deep convolution becomes more and more vivid as the iteration number increases.
FIG. 7 is a graph of the change in crack detection accuracy before and after data set expansion. The X axis represents a virtual crack image generated by a generation type countermeasure network, the Y axis represents the average accuracy of crack detection, and the orange curve represents a crack detection accuracy rate change curve of a Faster R-CNN detection model after 1000 real crack image data sets are expanded. It can be seen from the graph that as the number of virtual crack images gradually increases, the crack detection accuracy is also continuously improved. Tests show that under the condition that the number of real crack images is certain, when a generative confrontation network is used for amplifying an original training set, the crack detection accuracy of a crack detection model can be improved to a certain extent.
Claims (6)
1. A method for establishing a virtual augmentation model of a pavement crack image is characterized by comprising the following steps:
step 1: collecting a pavement crack image, and preprocessing the pavement crack image to obtain a real pavement crack image set, wherein the pavement crack is a transverse seam or a longitudinal seam;
Step 2: establishing a DCGAN generation confrontation network model, wherein the DCGAN generation confrontation network model comprises a generator network and a discriminator network, and penalty terms are set behind loss functions of the generator network and the discriminator network;
and step 3: and (3) acquiring random noise, inputting the random noise and the real pavement crack image set acquired in the step (1) into the DCGAN generation countermeasure network model acquired in the step (2) for training, wherein the generator network model after training is the pavement crack image virtual augmentation model.
2. The method for establishing the virtual augmented model of the pavement crack image as claimed in claim 1, wherein the generator network has 5 layers, the first layer is a full-connected layer, the second to fifth layers are convolutional layers, and the convolutional layers all adopt Relu as an activation function.
3. The method for building a virtual augmented model of a pavement crack image as claimed in claim 1, wherein the discriminator network has 5 layers, the first to fourth layers are convolutional layers, the fifth layer is a fully-connected layer, and the convolutional layers all adopt leakyrelu as an activation function.
4. The method for establishing a virtual augmented model of a pavement crack image according to claim 1, wherein the loss function after adding a penalty term is 
And is
Wherein X represents the input of the generator network and the arbiter network and y represents the input of the generator network and the arbiter networkOutputting, wherein w is a weight vector, alpha represents that the value range of the regularization coefficient is 0 to 1, | w | | calculation1Is the sum of the absolute values of the weights w in the model.
5. The method for establishing the virtual augmented model of the pavement crack image as claimed in claim 1, wherein the preprocessing in the step 1 comprises the operations of decoloring, denoising, contrast enhancement, brightness enhancement and segmentation into uniform size which are sequentially performed.
6. An image virtual augmentation method comprising the steps of:
acquiring random noise, inputting the random noise into the pavement crack image virtual augmentation model obtained by the pavement crack image virtual augmentation model establishing method according to any one of claims 1 to 5, and obtaining a pavement crack image augmentation data set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010574126.5A CN111861906B (en) | 2020-06-22 | 2020-06-22 | Pavement crack image virtual augmentation model establishment and image virtual augmentation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010574126.5A CN111861906B (en) | 2020-06-22 | 2020-06-22 | Pavement crack image virtual augmentation model establishment and image virtual augmentation method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111861906A true CN111861906A (en) | 2020-10-30 |
| CN111861906B CN111861906B (en) | 2023-10-31 |
Family
ID=72987142
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010574126.5A Active CN111861906B (en) | 2020-06-22 | 2020-06-22 | Pavement crack image virtual augmentation model establishment and image virtual augmentation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111861906B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112396110A (en) * | 2020-11-20 | 2021-02-23 | 南京大学 | Method for generating anti-cascade network augmented image |
| CN112415514A (en) * | 2020-11-16 | 2021-02-26 | 北京环境特性研究所 | Target SAR image generation method and device |
| CN112862706A (en) * | 2021-01-26 | 2021-05-28 | 北京邮电大学 | Pavement crack image preprocessing method and device, electronic equipment and storage medium |
| CN113009447A (en) * | 2021-03-05 | 2021-06-22 | 长安大学 | Road underground cavity detection early warning method based on deep learning and ground penetrating radar |
| CN113222114A (en) * | 2021-04-22 | 2021-08-06 | 北京科技大学 | Image data augmentation method and device |
| CN113592000A (en) * | 2021-08-03 | 2021-11-02 | 成都理工大学 | Convolution-based crack identification technology for antagonistic neural network |
| CN113822880A (en) * | 2021-11-22 | 2021-12-21 | 中南大学 | Crack identification method based on deep learning |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018028255A1 (en) * | 2016-08-11 | 2018-02-15 | 深圳市未来媒体技术研究院 | Image saliency detection method based on adversarial network |
| CN108876780A (en) * | 2018-06-26 | 2018-11-23 | 陕西师范大学 | Bridge Crack image crack detection method under a kind of complex background |
| CN109345507A (en) * | 2018-08-24 | 2019-02-15 | 河海大学 | A kind of dam image crack detection method based on transfer learning |
| CN110956591A (en) * | 2019-11-06 | 2020-04-03 | 河海大学 | Dam crack image data enhancement method based on depth convolution generation countermeasure network |
-
2020
- 2020-06-22 CN CN202010574126.5A patent/CN111861906B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018028255A1 (en) * | 2016-08-11 | 2018-02-15 | 深圳市未来媒体技术研究院 | Image saliency detection method based on adversarial network |
| CN108876780A (en) * | 2018-06-26 | 2018-11-23 | 陕西师范大学 | Bridge Crack image crack detection method under a kind of complex background |
| CN109345507A (en) * | 2018-08-24 | 2019-02-15 | 河海大学 | A kind of dam image crack detection method based on transfer learning |
| CN110956591A (en) * | 2019-11-06 | 2020-04-03 | 河海大学 | Dam crack image data enhancement method based on depth convolution generation countermeasure network |
Non-Patent Citations (2)
| Title |
|---|
| 李良福;孙瑞;: "复杂背景下基于图像处理的桥梁裂缝检测算法", 激光与光电子学进展, no. 06 * |
| 胡敏;李良福;: "基于生成式对抗网络的裂缝图像修复方法", 计算机应用与软件, no. 06 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112415514A (en) * | 2020-11-16 | 2021-02-26 | 北京环境特性研究所 | Target SAR image generation method and device |
| CN112415514B (en) * | 2020-11-16 | 2023-05-02 | 北京环境特性研究所 | Target SAR image generation method and device |
| CN112396110A (en) * | 2020-11-20 | 2021-02-23 | 南京大学 | Method for generating anti-cascade network augmented image |
| CN112396110B (en) * | 2020-11-20 | 2024-02-02 | 南京大学 | Method for generating augmented image of countermeasure cascade network |
| CN112862706A (en) * | 2021-01-26 | 2021-05-28 | 北京邮电大学 | Pavement crack image preprocessing method and device, electronic equipment and storage medium |
| CN113009447A (en) * | 2021-03-05 | 2021-06-22 | 长安大学 | Road underground cavity detection early warning method based on deep learning and ground penetrating radar |
| CN113222114A (en) * | 2021-04-22 | 2021-08-06 | 北京科技大学 | Image data augmentation method and device |
| CN113222114B (en) * | 2021-04-22 | 2023-08-15 | 北京科技大学 | Image data augmentation method and device |
| CN113592000A (en) * | 2021-08-03 | 2021-11-02 | 成都理工大学 | Convolution-based crack identification technology for antagonistic neural network |
| CN113822880A (en) * | 2021-11-22 | 2021-12-21 | 中南大学 | Crack identification method based on deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111861906B (en) | 2023-10-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111861906A (en) | Pavement crack image virtual augmentation model establishment and image virtual augmentation method | |
| CN109345507B (en) | Dam image crack detection method based on transfer learning | |
| CN113673307B (en) | Lightweight video action recognition method | |
| CN110188685B (en) | Target counting method and system based on double-attention multi-scale cascade network | |
| CN109493303B (en) | Image defogging method based on generation countermeasure network | |
| CN109035149B (en) | License plate image motion blur removing method based on deep learning | |
| CN113313657B (en) | Unsupervised learning method and system for low-illumination image enhancement | |
| CN108711141B (en) | Motion blurred image blind restoration method using improved generation type countermeasure network | |
| CN112184577B (en) | Single image defogging method based on multiscale self-attention generation countermeasure network | |
| CN106228185A (en) | A kind of general image classifying and identifying system based on neutral net and method | |
| CN112489164B (en) | Image coloring method based on improved depth separable convolutional neural network | |
| CN111914728B (en) | Hyperspectral remote sensing image semi-supervised classification method and device and storage medium | |
| CN111046917B (en) | Object-based enhanced target detection method based on deep neural network | |
| CN109461458B (en) | Audio anomaly detection method based on generation countermeasure network | |
| CN112819096B (en) | Construction method of fossil image classification model based on composite convolutional neural network | |
| CN110648292A (en) | High-noise image denoising method based on deep convolutional network | |
| CN108764250A (en) | A method of extracting essential image with convolutional neural networks | |
| CN113538234A (en) | Remote sensing image super-resolution reconstruction method based on lightweight generation model | |
| CN113870124A (en) | Dual-network mutual excitation learning shadow removing method based on weak supervision | |
| CN114581552A (en) | Gray level image colorizing method based on generation countermeasure network | |
| CN114463176B (en) | Image super-resolution reconstruction method based on improved ESRGAN | |
| CN111815526A (en) | Rain image rainstrip removing method and system based on image filtering and CNN | |
| CN113989612A (en) | Remote sensing image target detection method based on attention and generation countermeasure network | |
| CN112906591A (en) | Radar radiation source identification method based on multi-stage jumper residual error network | |
| CN115860113B (en) | Training method and related device for self-countermeasure neural network model |
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 |