CN111323764B - Underground engineering target body intelligent identification method and system based on ground penetrating radar - Google Patents
Underground engineering target body intelligent identification method and system based on ground penetrating radar Download PDFInfo
- Publication number
- CN111323764B CN111323764B CN202010071046.8A CN202010071046A CN111323764B CN 111323764 B CN111323764 B CN 111323764B CN 202010071046 A CN202010071046 A CN 202010071046A CN 111323764 B CN111323764 B CN 111323764B
- Authority
- CN
- China
- Prior art keywords
- ground penetrating
- penetrating radar
- target body
- data
- underground engineering
- 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
Images
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
- 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
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
-
- 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
Abstract
The invention discloses a ground penetrating radar-based underground engineering target body intelligent identification method and a system, wherein the method comprises the following steps: acquiring a computer simulation data set, wherein the data set comprises a plurality of groups of data pairs of a ground penetrating radar data profile diagram and a target body tag diagram; training to obtain a target body recognition model based on the data set by adopting a deep learning network; and acquiring radar detection data acquired in real time, and identifying the target body by adopting the target body identification model. The ground penetrating radar data real-time identification and output method can realize real-time identification and output of the target body aiming at the ground penetrating radar data.
Description
Technical Field
The invention relates to the technical field of geological detection technology and artificial intelligence, in particular to an underground engineering target body intelligent identification method and system based on a ground penetrating radar.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
At present, the detection of the underground structure is mainly to analyze the data collected by the ground penetrating radar, the abnormity is found through the reflection of radar waves, and the judgment is carried out by using an artificial expert to explain what object the abnormity in the detection data of the ground penetrating radar belongs to or matching the abnormity with a database of the detection data of the ground penetrating radar, so that a large amount of manpower and material resources are consumed, and the detection efficiency is extremely low.
Although the deep learning method is applied to the identification of the ground penetrating radar detection data, the deep learning method mainly focuses on a specific research field, and the identification result is whether an abnormality exists or not, so that the position of the abnormality cannot be accurately given. Before the model is established, a large amount of ground penetrating radar detection data covering different abnormal types need to be collected in advance, but the coverage range of the abnormal types is very large, the data collection work is complicated, and the established model is often poor in universality.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides the underground engineering target body intelligent identification method and system based on the ground penetrating radar, which can be used for quickly and effectively identifying abnormal bodies, diseases and other target bodies aiming at the detection data of the ground penetrating radar.
In order to achieve the above object, one or more embodiments of the present invention provide the following technical solutions:
an underground engineering target body intelligent identification method based on a ground penetrating radar comprises the following steps:
acquiring a computer simulation data set, wherein the data set comprises a plurality of groups of data pairs of a ground penetrating radar data profile diagram and a target body tag diagram;
training to obtain a target body recognition model based on the data set by adopting a deep learning network;
and acquiring radar detection data acquired in real time, and identifying the target body by adopting the target body identification model.
Further, constructing the computer simulation data set includes:
simulating dielectric constant models of various underground engineering target bodies with different background media, shapes, sizes and distributions;
forward modeling is carried out on each dielectric constant model, radar data profile maps with different frequencies are respectively generated, and a plurality of groups of ground penetrating radar data profile map-target body dielectric constant model data pairs are obtained;
and carrying out contour recognition on the target body dielectric constant model in each group of ground penetrating radar data profile diagram-target body dielectric constant model data pairs, assigning the pixel values within the contour range to corresponding type identification codes, and obtaining a plurality of groups of ground penetrating radar data profile diagram-target body label diagram data pairs.
Further, the deep learning network model architecture comprises: the system comprises a time dimension compression network structure used for extracting features of the radar detection data profile, a self-coding network structure used for coding each feature channel, and a label graph decoding network structure used for decoding each feature channel.
Further, the time dimension compression network structure comprises a convolution structure formed by sequentially cascading 6 layers, wherein the output end of the 2 nd layer convolution structure is further connected with the input end of the residual block, and the output end of the residual block is connected to the input end of the 3 rd layer convolution structure.
Further, the self-coding network structure comprises 4 fully-connected layers.
Further, the tag graph code network structure comprises a convolution structure formed by sequentially cascading 6 layers, wherein the output end of the convolution structure of the 2 nd layer is further connected with the input end of the residual block, and the output end of the residual block is connected with the input end of the convolution structure of the 3 rd layer.
And further, acquiring a plurality of groups of real ground penetrating radar data profile diagram-target body label diagram data pairs, and performing parameter adjustment on the obtained target body identification model.
One or more embodiments provide an electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing the method for ground penetrating radar-based intelligent identification of a target entity of an underground project when executing the program.
One or more embodiments provide a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method for intelligent identification of a subsurface engineering target body based on a ground penetrating radar.
One or more embodiments provide a ground penetrating radar-based underground engineering target body identification system, comprising: the system comprises an integrated ground penetrating radar, an embedded control module and a touch screen; wherein the content of the first and second substances,
the touch screen receives the selection of a user about a ground penetrating radar acquisition mode and the setting of the number of data channels transmitted by the ground penetrating radar, generates a starting instruction and sends the starting instruction to the embedded control module to control the integrated ground penetrating radar to start detection work;
the embedded control module receives data acquired by the ground penetrating radar in real time, judges whether the data reaches the number of tracks set by a user, and continues to acquire the data if the data does not reach the set number of tracks; and if the number of the underground engineering targets reaches the number, carrying out target body identification by adopting the underground engineering target body intelligent identification method based on the ground penetrating radar, and outputting a detection result in real time.
The above one or more technical solutions have the following beneficial effects:
the target body identification model based on radar detection data is established, and real-time detection aiming at abnormal target bodies in the radar detection process can be realized.
The method acquires the radar detection image-target body label image data pair as training data in an analog simulation mode, acquires the radar detection profile image and the target body label image of underground engineering target bodies aiming at various types, background media, shapes, sizes and distributions, and ensures that the training data has enough data volume and also provides guarantee for the generalization capability of a subsequent model; and real data pairs are acquired after the model training and are used as test data, and parameters of the trained model are adjusted, so that the identification of a target body suitable for real environment radar detection data can be obtained.
When feature learning is carried out on radar detection data, a one-dimensional rolling machine is adopted for time dimension compression, the problem of unbalanced data volume of time dimension and space dimension is solved, the complete and accurate extraction of the target body reflection signal features is guaranteed, the guarantee is provided for the space corresponding learning of the subsequent features, and the model precision is guaranteed.
The invention also provides a target body identification system which can accurately identify the outline of the target body in the profile and can restore the shape of the underground engineering target body through continuous detection of the radar profile.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of the invention and together with the description serve to explain the invention and not to limit the invention.
Fig. 1 is a flowchart of a ground penetrating radar intelligent identification method according to an embodiment of the present invention;
FIG. 2 is a model of the subsurface structure of a simulated data set created in one embodiment of the present invention;
FIG. 3 is a ground penetrating radar detection data of a set of simulated data created in an embodiment of the present invention;
FIG. 4 is a time dimension compressed network structure constructed in one embodiment of the present invention;
FIG. 5 is a network structure for spatial distribution feature self-encoding and tag map decoding constructed in one embodiment of the present invention;
FIG. 6 is a schematic diagram of a ground penetrating radar intelligent recognition system according to an embodiment of the present invention;
FIG. 7 is a flowchart illustrating the operation of the ground penetrating radar intelligent recognition system according to an embodiment of the present invention;
FIGS. 8(a) -8 (c) are radar detection data, actual tag maps, and target detection results, respectively, according to an embodiment of the present invention;
FIGS. 9(a) -9 (c) are radar detection data, actual tag maps and target detection results, respectively, of yet another embodiment of the present invention;
fig. 10(a) -10 (c) are radar detection data, an actual tag map, and a target detection result, respectively, according to still another embodiment of the present invention.
Detailed Description
It is to be understood that the following detailed description is exemplary and is intended to provide further explanation of the invention as claimed. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
The embodiments and features of the embodiments of the present invention may be combined with each other without conflict.
Example one
The embodiment discloses an underground engineering target body intelligent identification method based on a ground penetrating radar, which comprises the following steps as shown in figure 1:
and step S1, acquiring a computer simulation data set, wherein the data set comprises a plurality of groups of data pairs of a ground penetrating radar data profile diagram and a target body label diagram.
The identification method of the embodiment is mainly used for hidden abnormal bodies or diseases, including geological abnormal bodies such as broken, karst caves and faults of underground engineering, concrete reinforcement and structural diseases, pavement diseases such as pavement cracks and settlement, and various underground pipelines. Each type object body is assigned with a unique type identification code.
Fig. 2 and 3 are a label map of a hidden target and a ground penetrating radar data profile in a computer simulation data set, respectively. The method for constructing a plurality of groups of data pairs of the ground penetrating radar data profile map and the target body tag map comprises the following steps:
simulating dielectric constant models of various underground engineering target bodies with different background media, shapes, sizes and distributions;
forward modeling is carried out on each dielectric constant model, radar data profile maps with different frequencies are respectively generated, and a plurality of groups of ground penetrating radar data profile map-target body dielectric constant model data pairs are obtained;
and carrying out contour recognition on the target body dielectric constant model in each group of ground penetrating radar data profile diagram-target body dielectric constant model data pairs, assigning the pixel values within the contour range to corresponding type identification codes, and obtaining a plurality of groups of ground penetrating radar data profile diagram-target body label diagram data pairs.
The label graph carries out pixel-level labeling on the target body according to the type of the target body, and outlines of different types of target bodies are distinguished by different pixel values.
S2, constructing a ground penetrating radar intelligent recognition deep learning network architecture;
the ground penetrating radar intelligent recognition deep learning network consists of 16 layers in total and comprises a 6-layer time dimension compressed network structure, a 4-layer space distribution characteristic self-coding network structure and a 6-layer label graph decoding network structure.
As shown in fig. 4, the present embodiment adopts a time dimension compression network structure for feature extraction of the ground penetrating radar detection data. The 6-layer time dimension compression network structure consists of convolution structures of 3 x 1 and 3 x 3 and residual blocks; wherein, each layer of convolution structure comprises two times of operation of the rolling machine (namely, two times of operation of the rolling machine), the first time of operation of the rolling machine adopts a 3 multiplied by 1 rolling machine, the step length is 2 compression time dimension, the second time of operation of the rolling machine adopts a 3 multiplied by 3 rolling machine, and the step length is 1 stable characteristic. The residual block concatenates the output of the layer 2 convolution with the layer 4 convolution, preventing the gradient from vanishing and solving the degradation problem of the network. As can be seen from fig. 2 and fig. 3, the radar reflection signals of the abnormal body are represented in a parabolic shape in the radar cross-sectional diagrams, and one parabola corresponds to one disease, and since the data volume of the radar cross-sectional diagram input to the deep neural network is usually much larger than the data volume of the radar cross-sectional diagram in the time dimension, the feature extraction is directly performed by using conventional winding machines (such as winding machines 3 × 3,5 × 5, and the like), and it is difficult to ensure the integrity of the feature extraction of the abnormal body. Therefore, a one-dimensional rolling machine is adopted for time dimension compression, the problem of unbalanced time dimension-space dimension data volume is solved, the complete extraction of the target body reflection signal characteristics is ensured, and the quantity of the target bodies in the target body label graph output as a model is consistent.
As shown in fig. 5, the self-encoder network structure composed of 4 layers of fully-connected layers and the label graph decoder network structure composed of convolution and residual blocks on 6 layers are adopted. And the 4-layer self-encoder encodes each characteristic channel respectively and automatically learns the spatial correspondence between the ground penetrating radar detection data profile in each characteristic channel and the target body in the target body label graph. The self-encoder network structure is used for independently extracting each layer of features extracted after time compression by adopting a full-connection structure, automatically learning the spatial correspondence of the abnormal body type and the radar emission signal aiming at each layer of features, generating a feature map with the same dimension as the size of the label map and facilitating decoding. The 6-layer decoder uses convolution with convolution kernel size 4 x 4 or 3 x 3 to up-sample the feature map output from the encoder, and the residual block connects the output of the 12 th layer convolution (the 2 nd layer convolution of the decoder) and the 14 th layer convolution (the 4 th layer convolution of the decoder) together to prevent gradient disappearance and solve the degradation problem of the network to construct the final identified label map.
And S3, training the deep learning model, and obtaining parameters of the deep learning model to obtain the target body recognition model.
The deep learning model is trained by adopting a data set simulated by a computer and using Focal local or Cross-enhancement local and Lovasz-Softmax local in combination as a Loss function, wherein the data set simulated by the computer is composed of ground penetrating radar data with various frequencies and corresponding hidden target body label maps, and the ground penetrating radar data and the corresponding hidden target body label maps are input into the deep learning model together to train the deep learning model.
And S4, acquiring a plurality of groups of real ground penetrating radar data profile diagram-target body label diagram data pairs, and finely adjusting model parameters to obtain the ground penetrating radar intelligent identification model suitable for real data.
And acquiring a plurality of groups of real ground penetrating radar data profile-target body label graph data pairs as a test set, specifically, a radar detection data profile of real detection, and labeling according to target bodies in the data profile to obtain a target body label graph. And respectively carrying out fine adjustment on model parameters trained by using different loss functions by adopting a small amount of real data to obtain two ground penetrating radar intelligent recognition models suitable for the real data, wherein the ground penetrating radar intelligent recognition models can carry out performance evaluation by adopting evaluation indexes such as a confusion matrix, PA, MPA, MIoU, FWIoU and the like.
Example two
A ground penetrating radar-based underground engineering target body identification system, as shown in fig. 6, comprising: the system comprises a host-antenna integrated ground penetrating radar, an embedded control module and a touch screen. The integrated ground penetrating radar is connected with the embedded control module, the touch screen is connected with the embedded control module, and the intelligent identification model of the ground penetrating radar is embedded in the embedded control module.
The integrated ground penetrating radar is connected with the embedded control module through two data transmission modes of Ethernet or wireless Wifi.
The touch screen is connected with the embedded control module through a USB (universal serial bus), so that bidirectional communication and man-machine interaction are realized.
The embedded control module is provided with an expansion communication interface, the PCIe interface is connected with the 4G module, and Wifi or 4G is adopted to transmit real-time ground penetrating radar detection data and identification results to electronic equipment such as a mobile phone end, a tablet computer and a computer, so that real-time display of an underground engineering structure is realized.
The touch screen is used for inputting start-stop control instructions, collection mode selection and channel number setting collected by the ground penetrating radar, and outputting and displaying ground penetrating radar detection data and identification results. The acquisition mode of the ground penetrating radar comprises a time mode and a distance mode, a user can select the two modes and set the channel number of data acquired by the ground penetrating radar, and the data detected by the ground penetrating radar reaching the set channel number are transmitted to the deep learning embedded control module to be identified.
As shown in fig. 7, the working method of the ground penetrating radar intelligent identification system is as follows:
step 1: the system is powered on, the embedded control module transmits and displays man-machine interaction instructions such as start-stop control, acquisition mode selection, channel number setting and the like to the touch screen, the touch screen receives user selection about a ground penetrating radar acquisition mode and the setting of the number of data channels transmitted by the ground penetrating radar, a starting instruction is generated and sent to the embedded control module, and the integrated ground penetrating radar is controlled to start detection work;
step 2: the embedded control module receives data collected by the ground penetrating radar in real time, judges whether the number of channels set by a user is reached, and transmits the data to the intelligent recognition model of the ground penetrating radar if the number of channels is reached; if the set number of tracks is not reached, continuing to collect the data until the collected data reaches the set number of tracks, and then transmitting the data;
and step 3: the embedded control module performs target body detection on the received data through the ground penetrating radar intelligent identification model and outputs a detection result in real time; and when the target is detected, generating a target label graph, outputting and displaying the target label graph and corresponding ground penetrating radar detection data, and alarming. Fig. 8(a) -8 (c), 9(a) -9 (c), and 10(a) -10 (c) are schematic diagrams of detection results for three radar detection data cross-sectional views, respectively.
The ground penetrating radar detection data and the identified tag map can be transmitted to a touch screen for displaying, and can also be transmitted to electronic equipment such as a mobile phone terminal, a tablet personal computer and a computer in real time through Wifi or 4G, so that the real-time display of the underground engineering structure is realized. And finally, judging whether the data acquisition of the ground penetrating radar is finished or not, if not, continuing the process, and finishing the work of the intelligent recognition system of the ground penetrating radar after finishing the data acquisition.
The embedded control module adopts a high-performance and low-power-consumption deep learning embedded main control module NVIDIA Netson based on NVIDIA Pascal architectureTM. The embedded operating system has the characteristics of high performance, small volume, low power consumption and the like, can be loaded with the Ubuntu system and embedded with a deep learning algorithm, directly runs a neural network based on a PyTorch frame, and controls the acquisition of ground penetrating radar detection data, the display of ground penetrating radar detection data and an identified label graph and the transmission of the data.
In step 3, the method for identifying a target body specifically refers to the relevant description part of the first embodiment.
EXAMPLE III
The embodiment aims at providing an electronic device.
An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing the following steps when executing the program, comprising:
acquiring a computer simulation data set, wherein the data set comprises a plurality of groups of data pairs of a ground penetrating radar data profile diagram and a target body tag diagram;
training to obtain a target body recognition model based on the data set by adopting a deep learning network;
and acquiring radar detection data acquired in real time, and identifying the target body by adopting the target body identification model.
Example four
An object of the present embodiment is to provide a computer-readable storage medium.
A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, performs the steps of:
acquiring a computer simulation data set, wherein the data set comprises a plurality of groups of data pairs of a ground penetrating radar data profile diagram and a target body tag diagram;
training to obtain a target body recognition model based on the data set by adopting a deep learning network;
and acquiring radar detection data acquired in real time, and identifying the target body by adopting the target body identification model.
The specific implementation manner of each step in the third to fourth embodiments is referred to in the relevant description part of the first embodiment.
One or more of the above embodiments have the following technical effects:
a target body identification model based on radar detection data is established, and real-time detection aiming at abnormal target bodies in the radar detection process can be realized.
Acquiring a radar detection map-target body label map data pair as training data in an analog simulation mode, and acquiring a radar detection profile map and a target body label map of underground engineering targets with various types, background media, shapes, sizes and distributions, so that the training data has enough data volume and the generalization capability of a subsequent model is guaranteed; and real data pairs are acquired after the model training and are used as test data, and parameters of the trained model are adjusted, so that the identification of a target body suitable for real environment radar detection data can be obtained.
When feature learning is carried out on radar detection data, a one-dimensional rolling machine is adopted for time dimension compression, the problem of unbalanced data volume of time dimension and space dimension is solved, the complete and accurate extraction of the target body reflection signal features is guaranteed, the guarantee is provided for the space corresponding learning of the subsequent features, and the model precision is guaranteed.
The target body identification system can accurately identify the outline of a target body in the profile, and can restore the shape of the underground engineering target body through continuous detection of the radar profile.
Those skilled in the art will appreciate that the modules or steps of the present invention described above can be implemented using general purpose computer means, or alternatively, they can be implemented using program code that is executable by computing means, such that they are stored in memory means for execution by the computing means, or they are separately fabricated into individual integrated circuit modules, or multiple modules or steps of them are fabricated into a single integrated circuit module. The present invention is not limited to any specific combination of hardware and software.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the present invention, and it should be understood by those skilled in the art that various modifications and variations can be made without inventive efforts by those skilled in the art based on the technical solution of the present invention.
Claims (9)
1. An underground engineering target body intelligent identification method based on a ground penetrating radar is characterized by comprising the following steps:
acquiring a computer simulation data set, wherein the data set comprises a plurality of groups of data pairs of a ground penetrating radar data profile diagram and a target body tag diagram;
constructing a computer simulation dataset includes:
simulating dielectric constant models of underground engineering target bodies with various background media, shapes, sizes and distributions;
forward modeling is carried out on each dielectric constant model, radar data profile maps with different frequencies are respectively generated, and a plurality of groups of ground penetrating radar data profile map-target body dielectric constant model data pairs are obtained;
carrying out contour recognition on a target body dielectric constant model in each group of ground penetrating radar data profile diagram-target body dielectric constant model data pairs, assigning pixel values within a contour range to corresponding type identification codes, and obtaining a plurality of groups of ground penetrating radar data profile diagram-target body label diagram data pairs;
training to obtain a target body recognition model based on the data set by adopting a deep learning network;
and acquiring radar detection data acquired in real time, and identifying the target body by adopting the target body identification model.
2. The method for intelligently identifying underground engineering target bodies based on the ground penetrating radar as claimed in claim 1, wherein the deep learning network model architecture comprises: the system comprises a time dimension compression network structure used for extracting features of the radar detection data profile, a self-coding network structure used for coding each feature channel, and a label graph decoding network structure used for decoding each feature channel.
3. The method for intelligently identifying the underground engineering target body based on the ground penetrating radar as claimed in claim 2, wherein the time dimension compression network structure comprises a convolution structure formed by sequentially cascading 6 layers, wherein the output end of the convolution structure of the 2 nd layer is further connected with the input end of a residual block, and the output end of the residual block is connected with the input end of the convolution structure of the 4 th layer.
4. The method for intelligently identifying underground engineering target bodies based on ground penetrating radar as claimed in claim 2, wherein said self-coding network structure comprises 4 fully-connected layers.
5. The method for intelligently identifying underground engineering target bodies based on the ground penetrating radar as claimed in claim 2, wherein the tag diagram code network structure comprises 6 layers of convolution structures which are sequentially cascaded, wherein the output end of the 2 nd layer of convolution structure is also connected with the input end of a residual block, and the output end of the residual block is connected with the input end of the 4 th layer of convolution structure.
6. The method for intelligently identifying the underground engineering target body based on the ground penetrating radar as claimed in claim 1, wherein a plurality of groups of real data profile diagram-target body label diagram data pairs of the ground penetrating radar are obtained, and the obtained target body identification model is subjected to parameter adjustment to obtain the target body identification model suitable for practical engineering application.
7. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program implements a method for intelligent identification of a ground penetrating radar-based underground engineering target volume according to any one of claims 1 to 6.
8. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, implements a method for intelligent identification of a ground penetrating radar-based underground engineering target object according to any one of claims 1 to 6.
9. A ground penetrating radar-based underground engineering target body identification system is characterized by comprising: the system comprises an integrated ground penetrating radar, an embedded control module and a touch screen; wherein the content of the first and second substances,
the touch screen receives the selection of a user about a ground penetrating radar acquisition mode and the setting of the number of data channels transmitted by the ground penetrating radar, generates a starting instruction and sends the starting instruction to the embedded control module to control the integrated ground penetrating radar to start detection work;
the embedded control module receives data acquired by the ground penetrating radar in real time, judges whether the data reaches the number of tracks set by a user, and continues to acquire the data if the data does not reach the set number of tracks; if the number of the channels is reached, carrying out target body identification by adopting the underground engineering target body intelligent identification method based on the ground penetrating radar as claimed in any one of claims 1 to 6, and outputting the detection result in real time.
Priority Applications (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010071046.8A CN111323764B (en) | 2020-01-21 | 2020-01-21 | Underground engineering target body intelligent identification method and system based on ground penetrating radar |
| PCT/CN2020/119581 WO2021068848A1 (en) | 2019-10-09 | 2020-09-30 | Tunnel structure disease multi-scale measurement and intelligent diagnosis system and method |
| CN202080010321.3A CN113424055B (en) | 2019-10-09 | 2020-09-30 | Multi-scale tunnel structure disease detection and intelligent diagnosis system and method |
| US17/289,139 US20210396842A1 (en) | 2019-10-09 | 2020-09-30 | Multi-scale inspection and intelligent diagnosis system and method for tunnel structural defects |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010071046.8A CN111323764B (en) | 2020-01-21 | 2020-01-21 | Underground engineering target body intelligent identification method and system based on ground penetrating radar |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111323764A CN111323764A (en) | 2020-06-23 |
| CN111323764B true CN111323764B (en) | 2022-03-25 |
Family
ID=71167222
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010071046.8A Active CN111323764B (en) | 2019-10-09 | 2020-01-21 | Underground engineering target body intelligent identification method and system based on ground penetrating radar |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111323764B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021068848A1 (en) * | 2019-10-09 | 2021-04-15 | 山东大学 | Tunnel structure disease multi-scale measurement and intelligent diagnosis system and method |
| CN112232392B (en) * | 2020-09-29 | 2022-03-22 | 深圳安德空间技术有限公司 | Data interpretation and identification method for three-dimensional ground penetrating radar |
| CN112731381A (en) * | 2020-12-16 | 2021-04-30 | 华南农业大学 | Method for intelligently detecting hard foreign matters in soil by utilizing android debugging bridge and vehicle-mounted radar |
| CN113447536B (en) * | 2021-06-24 | 2022-09-30 | 山东大学 | Concrete dielectric constant inversion and disease identification method and system |
| JP2022142602A (en) * | 2021-03-16 | 2022-09-30 | 独立行政法人国立高等専門学校機構 | Electromagnetic wave rader device and method for learning electromagnetic wave rader device |
| CN115112854A (en) * | 2022-05-26 | 2022-09-27 | 北京理工大学 | Ground penetrating radar corrosion steel bar identification method based on polarization characteristics and machine learning |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104267440A (en) * | 2014-09-29 | 2015-01-07 | 中国神华能源股份有限公司 | Common middle point (CMP) detection method used for ground penetrating radar (GPR) |
| CN108182450A (en) * | 2017-12-25 | 2018-06-19 | 电子科技大学 | A kind of airborne Ground Penetrating Radar target identification method based on depth convolutional network |
| CN108830331A (en) * | 2018-06-22 | 2018-11-16 | 西安交通大学 | A kind of Ground Penetrating Radar object detection method based on full convolutional network |
| CN110458129A (en) * | 2019-08-16 | 2019-11-15 | 电子科技大学 | Nonmetallic mine recognition methods based on depth convolutional neural networks |
-
2020
- 2020-01-21 CN CN202010071046.8A patent/CN111323764B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104267440A (en) * | 2014-09-29 | 2015-01-07 | 中国神华能源股份有限公司 | Common middle point (CMP) detection method used for ground penetrating radar (GPR) |
| CN108182450A (en) * | 2017-12-25 | 2018-06-19 | 电子科技大学 | A kind of airborne Ground Penetrating Radar target identification method based on depth convolutional network |
| CN108830331A (en) * | 2018-06-22 | 2018-11-16 | 西安交通大学 | A kind of Ground Penetrating Radar object detection method based on full convolutional network |
| CN110458129A (en) * | 2019-08-16 | 2019-11-15 | 电子科技大学 | Nonmetallic mine recognition methods based on depth convolutional neural networks |
Non-Patent Citations (1)
| Title |
|---|
| 基于BP神经网络的探地雷达图像特征判识与提取研究;温世儒等;《公路》;20180731(第7期);第312-317页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111323764A (en) | 2020-06-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111323764B (en) | Underground engineering target body intelligent identification method and system based on ground penetrating radar | |
| CN111781576A (en) | Ground penetrating radar intelligent inversion method based on deep learning | |
| CN109493320B (en) | Remote sensing image road extraction method and system based on deep learning, storage medium and electronic equipment | |
| CN110348288A (en) | A kind of gesture identification method based on 77GHz MMW RADAR SIGNAL USING | |
| CN113009447B (en) | Road underground cavity detection and early warning method based on deep learning and ground penetrating radar | |
| CN110827406A (en) | Method for realizing rapid three-dimensional reconstruction of large water network tunnel | |
| CN110672072A (en) | Information multivariate interpretation and characterization method for fractured rock mass structure | |
| CN114332977A (en) | Key point detection method and device, electronic equipment and storage medium | |
| CN114022825A (en) | Precast beam field pedestal state recognition device based on deep learning and recognition method thereof | |
| CN115984273A (en) | Road disease detection method and device, computer equipment and readable storage medium | |
| CN117292148B (en) | Tunnel surrounding rock level assessment method based on directional drilling and test data | |
| Yao et al. | Application of improved MobileNet-SSD on underwater sea cucumber detection robot | |
| CN117454495A (en) | CAD vector model generation method and device based on building sketch outline sequence | |
| CN110163201B (en) | Image testing method and device, storage medium and electronic device | |
| CN105096334B (en) | A kind of mining area monitoring method and system | |
| CN113447536B (en) | Concrete dielectric constant inversion and disease identification method and system | |
| CN111062917A (en) | Depth neural network-based three-linear array image cloud detection method and device | |
| CN115565082A (en) | Method, system and device for removing cloud noise of satellite remote sensing image | |
| CN212459378U (en) | Ground object hyperspectral meter remote sensing land utilization sample acquisition instrument based on unmanned aerial vehicle | |
| Lu et al. | A novel global-aware deep network for road detection of very high resolution remote sensing imagery | |
| CN113838078A (en) | Coal mining subsidence ground crack identification and extraction method, device and storage medium | |
| CN112153347A (en) | Coal mine underground intelligent visual perception terminal, perception method, storage medium and electronic equipment | |
| Won et al. | Multi-task deep learning design and training tool for unified visual driving scene understanding | |
| CN204695400U (en) | A kind of pipeline construction figure automatic creation system | |
| Ding | A Transformer-Based Framework for Misfire Detection From Blasting-Induced Ground Vibration Signal |
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 |