CN114882236B - Automatic identification method for underground cavity target of ground penetrating radar based on SinGAN algorithm - Google Patents
Automatic identification method for underground cavity target of ground penetrating radar based on SinGAN algorithm Download PDFInfo
- Publication number
- CN114882236B CN114882236B CN202110542902.8A CN202110542902A CN114882236B CN 114882236 B CN114882236 B CN 114882236B CN 202110542902 A CN202110542902 A CN 202110542902A CN 114882236 B CN114882236 B CN 114882236B
- Authority
- CN
- China
- Prior art keywords
- ground penetrating
- penetrating radar
- image
- algorithm
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000000149 penetrating effect Effects 0.000 title claims abstract description 67
- 238000000034 method Methods 0.000 title claims abstract description 21
- 238000012549 training Methods 0.000 claims abstract description 33
- 238000013135 deep learning Methods 0.000 claims abstract description 21
- 238000001514 detection method Methods 0.000 claims abstract description 20
- 238000002372 labelling Methods 0.000 claims abstract description 12
- 238000012545 processing Methods 0.000 claims abstract description 10
- 238000012795 verification Methods 0.000 claims abstract description 10
- 230000003416 augmentation Effects 0.000 claims abstract description 9
- 208000009119 Giant Axonal Neuropathy Diseases 0.000 claims description 17
- 201000003382 giant axonal neuropathy 1 Diseases 0.000 claims description 17
- 230000001502 supplementing effect Effects 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 7
- 241000940966 Dexia Species 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000002910 structure generation Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/885—Radar or analogous systems specially adapted for specific applications for ground probing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Electromagnetism (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention provides a method for automatically identifying underground cavity targets of a ground penetrating radar based on SinGAN algorithm. Step 1: performing image augmentation processing on the obtained ground penetrating radar echo image of the underground cavity target by utilizing SinGAN algorithm to obtain a processed ground penetrating radar echo image with similar distribution; step 2: labeling the echo image of the ground penetrating radar generated in the step 1 to determine the position of a relevant target pixel; step 3: randomly distributing the marked data in the step 2 to a training set and a verification set; step 4: training a deep learning target recognition algorithm by using the training set and the verification set in the step 3 to obtain a weight model; step 5: and (3) inputting the weight model obtained in the step (4) into an existing deep learning algorithm model, and carrying out target identification detection on the echo image of the underground cavity target ground penetrating radar. The method solves the problem that the existing method is difficult to detect and identify the underground cavity target.
Description
Technical Field
The invention belongs to the field of target detection of echo diagram post-processing of ground penetrating radar; in particular to a method for automatically identifying underground cavity targets of a ground penetrating radar based on SinGAN algorithm.
Background
Ground penetrating radar is a non-invasive detection instrument for detecting shallow underground environments. The ground penetrating radar utilizes the difference of electromagnetic dielectric constants of underground mediums, the difference of the parameters is embodied in radar echo data, and the distribution of the underground environment can be rapidly detected and intuitively understood through processing the echo data. For visual presentation of echo data for manual analysis, it is a common method to list multi-channel echo data laterally, from which B-Scan images are commonly obtained in ground penetrating radar analysis.
The ground penetrating radar is an important geophysical method for rapid, high-resolution and nondestructive detection, and has important significance and value in underground collapse cavity detection research and engineering practice. The ground penetrating radar technology can not generate structural damage to the road surface, is suitable for various road conditions, has real-time and high accuracy in detection results, meets the requirements of road disease detection on high efficiency, no damage, accuracy and wide application range, and is suitable for detecting underground cavities of the road. The ground penetrating radar system can be composed of one or more pairs of transmitting and receiving antennas, each pair of transmitters and receivers can acquire a single B-Scan image by scanning a region of interest, and the distribution situation of the underground environment can be obtained through analysis and verification of the B-Scan image. The B-Scan image acquired in the actual engineering at present needs to be interpreted and interpreted manually, and the method has low efficiency and often causes the problem of missed detection or false detection. The detection and identification of underground cavity targets by using some currently mainstream deep learning methods also have problems that the underground cavity acquisition of the related mode information is difficult through confirmation, verification, positioning and acquisition, and the underground cavity has no fixed mode and shape in the B-Scan image, so that the acquisition of a large number of underground cavity samples is a difficult engineering task.
Disclosure of Invention
The invention provides an automatic identification method of underground cavity targets of a ground penetrating radar based on SinGAN algorithm, which is used for solving the problem that the existing method is difficult to detect and identify underground cavity targets.
The invention is realized by the following technical scheme:
an automatic identification method of underground cavity targets of a ground penetrating radar based on SinGAN algorithm comprises the following steps:
Step 1: performing image augmentation processing on the obtained ground penetrating radar echo image of the underground cavity target by utilizing SinGAN algorithm to obtain a processed ground penetrating radar echo image with similar distribution;
Step 2: labeling the echo image of the ground penetrating radar generated in the step 1 to determine the position of a relevant target pixel;
step 3: randomly distributing the marked data in the step 2 to a training set and a verification set;
Step 4: training a deep learning target recognition algorithm by using the training set and the verification set in the step 3 to obtain a weight model;
step 5: and (3) inputting the weight model obtained in the step (4) into an existing deep learning algorithm model, and carrying out target identification detection on the echo image of the underground cavity target ground penetrating radar.
Further, the step 1 uses SinGAN algorithm to perform image augmentation processing on the obtained ground penetrating radar echo image of the underground cavity target, specifically, the SinGAN algorithm includes n GANs with different scales, G n～G0 is respectively used, the scales are sequentially increased, the input of the G n th generator is a noise image, the output result of training is close to the original image x n after downsampling, the input of each GAN is the result output by G n+1 plus a noise image, except G n, the detail information of the last GAN output result is learned by each GAN;
From the above, sinGAN algorithm consists of a pyramid structure { G 0,...,GN }, for x: { x0,..x n } image pyramid training, where x n is a downsampled version of x, with a multiplier r n, for some r > 1, each generator G n is responsible for generating a true image sample w.r.t., the color block distribution in the corresponding image x n;
From the above, it can be seen that all generators and discriminators have the same receiving domain, and thus the size of the structure captured during the generation is reduced; g n maps spatial Gaussian white noise z N to image samples I.e.
In addition to the spatial noise z n, each generator G n also accepts an up-sampled version of the coarser scale image, i.e
G n execute operation
Wherein ψ n is a convolution block with 5 Conv (3 x 3) -BatchNorm-LeakyReLu; the training penalty for the nth GAN includes a resistance formula, i.e
Furthermore, labeling the relevant target pixel positions in the step 2 is specifically that the processed ground penetrating radar echo images with similar distribution are obtained, namely, labeling the relevant pixel positions of the generated ground penetrating radar echo images by using a labelimg tool, and determining whether each sub-pixel belongs to a cavity label or not and performing global labeling.
Further, the training of the deep learning target recognition algorithm in step 4 is performed to obtain a weight model, specifically, training the deep learning target recognition algorithm YOLOv, with training parameters of 16batch and 1000epochs, and finally obtaining a trained weight model.
Further, the step 5 of performing target recognition detection on the underground cavity target ground penetrating radar echo image specifically includes inputting the underground cavity target ground penetrating radar echo image which is not input into the system into the frame according to the trained deep learning algorithm model, automatically performing target recognition detection on the underground cavity target ground penetrating radar echo image, and finally outputting the processed image, wherein the processed image comprises the suspected underground cavity target selected by the frame selection frame and the confidence level thereof.
The beneficial effects of the invention are as follows:
According to the method, the ground penetrating radar echo map random generation work is carried out on the extracted features through the pyramid structure generation network, then the deep learning algorithm training and recognition are carried out according to the output result, the output result is used for realizing the ground penetrating radar echo image underground cavity target recognition, and the recognition probability can be effectively improved by detecting the underground cavity target of the ground penetrating radar echo map through the method.
The invention can improve the target recognition probability of the underground cavity to more than 90 percent.
In practice, when the ground penetrating radar collects data related to underground holes, the shape of the underground holes is random and difficult to predict, and meanwhile, the depth, the size and the position of the underground holes are unknown, so that the data collection and the subsequent classification and detection based on deep learning are greatly hindered. The invention aims to extract underground cavity target characteristics in a ground penetrating radar echo diagram by utilizing SinGAN image augmentation algorithm, randomly generate the ground penetrating radar echo diagram through pyramid structure generation network on the extracted characteristics, train and identify the ground penetrating radar echo diagram according to the output result by using deep learning algorithm, and use the output result for realizing the identification of the underground cavity target of the ground penetrating radar echo diagram.
Drawings
FIG. 1 is a flow chart of the automatic identification method of the underground cavity target of the ground penetrating radar based on SinGAN algorithm.
Fig. 2 is a block diagram of the generator and arbiter of SinGAN algorithm.
Fig. 3 is a ground penetrating radar echo image of an acquired single Zhang Dexia hole target.
Fig. 4 is a generated ground penetrating radar echo map with a similar distribution to a known underground cavity ground penetrating radar image.
Fig. 5 is an effect diagram of target recognition of an echo image of a ground penetrating radar of an underground cavity target.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
An automatic identification method of underground cavity targets of a ground penetrating radar based on SinGAN algorithm comprises the following steps:
Step 1: performing image augmentation processing on the obtained ground penetrating radar echo image of the underground cavity target by utilizing SinGAN algorithm to obtain a processed ground penetrating radar echo image with similar distribution;
Step 2: labeling the echo image of the ground penetrating radar generated in the step 1 to determine the position of a relevant target pixel;
Step 3: randomly distributing the marked data in the step 2 to a training set and a verification set; to increase the robustness of the overall system;
Step 4: training a deep learning target recognition algorithm by using the training set and the verification set in the step 3 to obtain a weight model;
step 5: and (3) inputting the weight model obtained in the step (4) into an existing deep learning algorithm model, and carrying out target identification detection on the echo image of the underground cavity target ground penetrating radar.
Further, the step 1 uses SinGAN algorithm to perform image augmentation processing on the obtained ground penetrating radar echo image of the underground cavity target, specifically, the SinGAN algorithm includes n GANs with different scales, G n～G0 is respectively used, the scales are sequentially increased, the input of the G n th generator is a noise image, the output result of training is close to the original image x n after downsampling, the input of each GAN is the result output by G n+1 plus a noise image, except G n, the detail information of the last GAN output result is learned by each GAN; this is accomplished through challenge training in which G n learns to fool an associated discriminator D n that attempts to distinguish a generated patch from a patch in x n;
From the above, sinGAN algorithm consists of a pyramid structure { G 0,...,GN }, for x: { x0,..x n } image pyramid training, where x n is a downsampled version of x, with a multiplier r n, for some r > 1, each generator G n is responsible for generating a true image sample w.r.t., the color block distribution in the corresponding image x n;
from the above, it can be seen that the generation of the image samples starts from the coarsest level, then passes through all the generators in turn, up to the finest level, and injects noise at each level; all generators and discriminators have the same receiving domain, so the size of the structure captured during generation is reduced; on a coarse scale, this generation is purely generated, G n maps spatial Gaussian white noise z N to image samples I.e.
This layer (which is superior to one that is composed of one generator and one arbiter, so this layer is worth the generator in the last natural segment) is typically half the image height, so G n will generate the overall layout of the image and the global structure of the object. The generator Gn (N < N) at each smaller scale adds detail that was not generated by the previous scale. Thus, in addition to the spatial noise z n, each generator G n also accepts an up-sampled version of the coarser scale image, i.e
All generators have a similar architecture as shown in fig. 2. Specifically, noise z n is added to the imageIs fed into a convolutional sequence layer. This ensures that the GAN does not ignore noise. Wherein the effect of the convolution layer is to generate missing detailsI.e. G n performs the operation
Wherein ψ n is a convolution block with 5 Conv (3 x 3) -BatchNorm-LeakyReLu; starting with 32 kernels per block on the coarsest scale, then increasing by a factor of 2 per 4 scales. Because the generator is fully convoluted, images of arbitrary size and aspect ratio can be generated at test time (by changing the size of the noise figure);
Training our multiscale architecture sequentially from the coarsest scale to the finer scale, once each GAN is trained, it is fixed; the training penalty for the nth GAN includes a resistance formula, i.e
Training a new unconditional generation model SinGAN on a single natural image, using a dedicated multi-scale challenge training scheme to learn patch statistics of images across multiple scales; this can then be used to generate a realistic new image sample that preserves the original patch tile distribution while creating new object configurations and structures.
Image augmentation processing is carried out on the obtained ground penetrating radar echo image (shown in fig. 3) of the single Zhang Dexia cavity target by utilizing SinGAN algorithm, and the processed ground penetrating radar echo image with similar distribution is shown in fig. 4.
Furthermore, labeling the relevant target pixel positions in the step 2 is specifically that the processed ground penetrating radar echo images with similar distribution are obtained, namely, labeling the relevant pixel positions of the generated ground penetrating radar echo images by using a labelimg tool, and determining whether each sub-pixel belongs to a cavity label or not and performing global labeling.
Further, the training of the deep learning target recognition algorithm in step 4 is performed to obtain a weight model, specifically, training the deep learning target recognition algorithm YOLOv, with training parameters of 16batch and 1000epochs, and finally obtaining a trained weight model.
Further, the step 5 of performing target recognition detection on the underground cavity target ground penetrating radar echo image specifically includes inputting the underground cavity target ground penetrating radar echo image which is not input into the system into the frame according to the trained deep learning algorithm model, automatically performing target recognition detection on the underground cavity target ground penetrating radar echo image, and finally outputting the processed image, wherein the processed image comprises the suspected underground cavity target selected by the frame selection frame and the confidence level thereof.
Claims (4)
1. The automatic identification method for the underground cavity target of the ground penetrating radar based on SinGAN algorithm is characterized by comprising the following steps:
Step 1: performing image augmentation processing on the obtained ground penetrating radar echo image of the underground cavity target by utilizing SinGAN algorithm to obtain a processed ground penetrating radar echo image with similar distribution;
Step 2: labeling the echo image of the ground penetrating radar generated in the step 1 to determine the position of a relevant target pixel;
step 3: randomly distributing the marked data in the step 2 to a training set and a verification set;
Step 4: training a deep learning target recognition algorithm by using the training set and the verification set in the step 3 to obtain a weight model;
step 5: inputting the weight model obtained in the step 4 into an existing deep learning algorithm model, and carrying out target identification detection on an echo image of the underground cavity target ground penetrating radar;
The step 1 is to perform image augmentation processing on the obtained ground penetrating radar echo image of the underground cavity target by using SinGAN algorithm, specifically, the SinGAN algorithm comprises n GANs with different scales, namely G n～G0, the scales are sequentially increased, the input of a G n generator is a noise image, the output result of training is close to the original image x n after downsampling, each GAN after the downsampling is input, the output result of G n+1 is added with a noise image, and except G n, each GAN is left to learn the detail information for supplementing the output result of the last GAN;
From the above, sinGAN algorithm consists of a pyramid structure { G 0,...,GN }, for x: { x0,..x n } image pyramid training, where x n is a downsampled version of x, with a multiplier r n, for some r > 1, each generator G n is responsible for generating a true image sample w.r.t., the color block distribution in the corresponding image x n;
From the above, it can be seen that all generators and discriminators have the same receiving domain, and thus the size of the structure captured during the generation is reduced; g n maps spatial Gaussian white noise z N to image samples I.e.
In addition to the spatial noise z n, each generator G n also accepts an up-sampled version of the coarser scale image, i.e
G n execute operation
Wherein ψ n is a convolution block with 5 Conv (3 x 3) -BatchNorm-LeakyReLu;
The training penalty for the nth GAN includes a resistance formula, i.e
2. The automatic identification method of the underground cavity target of the ground penetrating radar based on SinGAN algorithm according to claim 1, wherein the step 2 is characterized in that the labeling of the pixel positions of the relevant targets is specifically that the obtained ground penetrating radar echo images with similar distribution after processing is specifically that the correlation pixel position marking is carried out on the generated ground penetrating radar echo images by using labelimg tool, and whether each sub-pixel belongs to a cavity label or not is determined and global labeling is carried out.
3. The automatic identification method of the underground cavity target of the ground penetrating radar based on SinGAN algorithm according to claim 1, wherein the step 3 trains a deep learning target identification algorithm to obtain a weight model, specifically trains a deep learning target identification algorithm YOLOv5, and training parameters are 16batch and 1000epochs, and finally a trained weight model is obtained.
4. The automatic identification method for the underground cavity target of the ground penetrating radar based on SinGAN algorithm according to the claim 1 is characterized in that the step 5 is to carry out target identification detection on the underground cavity target ground penetrating radar echo image, specifically, according to the trained deep learning algorithm model, the newly acquired underground cavity target ground penetrating radar echo image is input into the SinGAN algorithm model, the target identification detection is automatically carried out on the underground cavity target ground penetrating radar echo image, and finally, the processed image is output, wherein the processed image comprises the suspected underground cavity target selected by the frame and the confidence level thereof.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110542902.8A CN114882236B (en) | 2021-05-19 | 2021-05-19 | Automatic identification method for underground cavity target of ground penetrating radar based on SinGAN algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110542902.8A CN114882236B (en) | 2021-05-19 | 2021-05-19 | Automatic identification method for underground cavity target of ground penetrating radar based on SinGAN algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114882236A CN114882236A (en) | 2022-08-09 |
| CN114882236B true CN114882236B (en) | 2024-09-13 |
Family
ID=82666736
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110542902.8A Active CN114882236B (en) | 2021-05-19 | 2021-05-19 | Automatic identification method for underground cavity target of ground penetrating radar based on SinGAN algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114882236B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112052899A (en) * | 2020-09-03 | 2020-12-08 | 上海无线电设备研究所 | Single ship target SAR image generation method based on generation countermeasure network |
| CN112462346A (en) * | 2020-11-26 | 2021-03-09 | 西安交通大学 | Ground penetrating radar roadbed defect target detection method based on convolutional neural network |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2019190998A (en) * | 2018-04-25 | 2019-10-31 | 株式会社日立情報通信エンジニアリング | Cavity identification system and cavity identification method |
| CN110472627B (en) * | 2019-07-02 | 2022-11-08 | 五邑大学 | End-to-end SAR image recognition method, device and storage medium |
| CN112666533B (en) * | 2020-12-31 | 2022-04-08 | 西安电子科技大学 | Repetition frequency change steady target identification method based on spatial pyramid pooling network |
-
2021
- 2021-05-19 CN CN202110542902.8A patent/CN114882236B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112052899A (en) * | 2020-09-03 | 2020-12-08 | 上海无线电设备研究所 | Single ship target SAR image generation method based on generation countermeasure network |
| CN112462346A (en) * | 2020-11-26 | 2021-03-09 | 西安交通大学 | Ground penetrating radar roadbed defect target detection method based on convolutional neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114882236A (en) | 2022-08-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107229918B (en) | SAR image target detection method based on full convolution neural network | |
| CN108416378B (en) | Large-scene SAR target recognition method based on deep neural network | |
| CN102609701B (en) | Remote sensing detection method based on optimal scale for high-resolution SAR (synthetic aperture radar) | |
| CN111476088B (en) | Asphalt pavement water damage identification model construction method, identification method and system | |
| CN103236063B (en) | Based on the SAR image oil spilling detection method of multiple dimensioned spectral clustering and decision level fusion | |
| CN102819740B (en) | A kind of Single Infrared Image Frame Dim targets detection and localization method | |
| CN112329542A (en) | SAR image ship target detection method based on feature refined network model | |
| CN111025285B (en) | Asphalt pavement water damage detection method based on map gray scale self-adaptive selection | |
| CN106446919A (en) | Quick detection method for ground penetrating radar hyperbolic curve target | |
| CN101533466B (en) | Image processing method for positioning eyes | |
| CN102208017A (en) | Ship detection method based on high-resolution synthetic aperture radar images | |
| CN103839075B (en) | SAR image classification method based on united sparse representation | |
| CN105787950A (en) | Infrared image sea-sky-line detection algorithm based on line gradient accumulation | |
| Changhui et al. | Cloud detection method based on feature extraction in remote sensing images | |
| CN114821358A (en) | Optical remote sensing image marine ship target extraction and identification method | |
| Xu et al. | The comparative study of three methods of remote sensing image change detection | |
| CN115499092A (en) | Astronomical radio transient signal searching method, system, device and readable storage medium | |
| CN107358162B (en) | Polarization SAR image change detection method based on depth stack network | |
| CN114882236B (en) | Automatic identification method for underground cavity target of ground penetrating radar based on SinGAN algorithm | |
| CN110223312A (en) | A kind of SAR image edge detection method based on shearing wave | |
| CN104732190B (en) | A kind of synthetic aperture sonar object detection method based on orthogonal texture correlation analysis | |
| CN116930904A (en) | Ground penetrating radar image alignment and difference detection method | |
| CN104077592B (en) | Automatic extraction method for high-resolution remote-sensing image navigation mark | |
| CN113723182B (en) | SAR image ship detection method under training sample limited condition | |
| CN103955915A (en) | SAR image segmentation based on sparse expression and multiple dictionaries |
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 |