CN114563785A - Earth surface deformation detection method, device, equipment and medium based on phase gradient - Google Patents
Earth surface deformation detection method, device, equipment and medium based on phase gradient Download PDFInfo
- Publication number
- CN114563785A CN114563785A CN202210092743.0A CN202210092743A CN114563785A CN 114563785 A CN114563785 A CN 114563785A CN 202210092743 A CN202210092743 A CN 202210092743A CN 114563785 A CN114563785 A CN 114563785A
- Authority
- CN
- China
- Prior art keywords
- phase gradient
- deformation
- neural network
- module
- unwrapped
- 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.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 51
- 238000000034 method Methods 0.000 claims abstract description 59
- 238000004364 calculation method Methods 0.000 claims abstract description 6
- 238000005305 interferometry Methods 0.000 claims abstract description 6
- 238000013528 artificial neural network Methods 0.000 claims description 32
- 238000012549 training Methods 0.000 claims description 23
- 230000007246 mechanism Effects 0.000 claims description 12
- 238000012360 testing method Methods 0.000 claims description 8
- 238000012795 verification Methods 0.000 claims description 6
- 208000037170 Delayed Emergence from Anesthesia Diseases 0.000 claims description 4
- 230000006870 function Effects 0.000 claims description 4
- 238000005259 measurement Methods 0.000 claims description 4
- 238000012300 Sequence Analysis Methods 0.000 abstract 1
- 238000012545 processing Methods 0.000 description 14
- 238000005516 engineering process Methods 0.000 description 11
- 238000012544 monitoring process Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 9
- 230000000694 effects Effects 0.000 description 7
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000004590 computer program Methods 0.000 description 5
- 238000013135 deep learning Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 5
- 238000004804 winding Methods 0.000 description 5
- 238000009825 accumulation Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 4
- 230000002708 enhancing effect Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 210000002569 neuron Anatomy 0.000 description 3
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 230000003287 optical effect Effects 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
- 238000004458 analytical method Methods 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003562 morphometric effect Effects 0.000 description 1
- 238000013425 morphometry Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 239000000523 sample Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
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
- G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
- G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
- G01S13/9004—SAR image acquisition techniques
-
- 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
- G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
- G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
- G01S13/9021—SAR image post-processing techniques
-
- 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
-
- 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)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Networks & Wireless Communication (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Electromagnetism (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The application discloses a method, a device, equipment and a medium for detecting earth surface deformation based on phase gradient, wherein the method comprises the following steps: obtaining a time sequence interferogram by adopting synthetic aperture radar interferometry; calculating a deformation phase gradient according to the time sequence interferogram; superposing the calculated deformation phase gradient, and performing unwrapping operation on the superposed deformation phase gradient to obtain an unwrapped absolute phase gradient; and inputting the unwrapped absolute phase gradient into a pre-trained local earth surface deformation identification model, and identifying a deformation area. According to the earth surface deformation detection method based on the phase gradient, the accuracy and the reliability of earth surface deformation detection are improved, the difficulty of identifying local deformation signals in a speed chart obtained by time sequence analysis is reduced, the detection efficiency of the local deformation signals in a large range is greatly improved, and meanwhile, the calculation cost is greatly reduced.
Description
Technical Field
The invention relates to the technical field of surface deformation detection, in particular to a surface deformation detection method, a surface deformation detection device, surface deformation detection equipment and a surface deformation detection medium based on phase gradients.
Background
The earth surface is influenced by natural factors and human activities, and can generate deformation with different degrees, which causes the earth surface buildings to be damaged, threatens the life safety of people and causes economic loss. The method can quickly and accurately detect the position of the ground surface deformation, and is of great importance to the prevention, control and treatment of ground deformation disasters and the health and safety evaluation of ground surface structure buildings. With the accumulation of synthetic aperture radar data, important observation support is provided for the detection of surface deformation. However, how to efficiently and quickly process massive SAR image data and acquire local deformation of the earth surface in a wide area still remains a challenge.
The synthetic aperture radar interferometry (InSAR for short) technology is a space geodetic measurement and remote sensing technology developed in recent decades, can obtain large-range ground surface deformation monitoring results, has the advantages of large monitoring range, high measuring precision, all weather all day and the like, and is widely used for monitoring ground surface deformation such as urban settlement, glacier movement, mining subsidence in mining areas, volcanic eruption, landslide, debris flow and the like. Over the years of development, the InSAR technology has gradually evolved from the traditional differential InSAR technology to the time-series InSAR technology. However, when the SAR image is affected by various loss-of-coherence factors, no matter the differential SAR or the multi-sequence SAR technology, under the conditions of time, long space baseline, vegetation coverage area and the like, the problems of low interference phase quality and poor interference effect inevitably occur, and further the number of time sequence monitoring points is rare, the monitoring precision is reduced, and the extraction of the earth surface deformation information in a wide area range is difficult, so that the earth surface deformation rule cannot be comprehensively revealed, and the application of the SAR technology in earth surface deformation monitoring is limited.
Disclosure of Invention
The embodiment of the application provides a method, a device, equipment and a medium for detecting earth surface deformation based on phase gradient. The following presents a simplified summary in order to provide a basic understanding of some aspects of the disclosed embodiments. This summary is not an extensive overview and is intended to neither identify key/critical elements nor delineate the scope of such embodiments. Its sole purpose is to present some concepts in a simplified form as a prelude to the more detailed description that is presented later.
In a first aspect, an embodiment of the present application provides a method for detecting surface deformation based on a phase gradient, including:
obtaining a time sequence interferogram by adopting synthetic aperture radar interferometry;
calculating a deformation phase gradient according to the time sequence interferogram;
superposing the calculated deformation phase gradient, and performing unwrapping operation on the superposed deformation phase gradient to obtain an unwrapped absolute phase gradient;
and inputting the unwrapped absolute phase gradient into a pre-trained local earth surface deformation identification model to obtain an identified deformation area.
In an alternative embodiment, computing the morphable phase gradient from the time series interferogram includes:
And calculating the deformation phase gradient of the azimuth direction and the deformation phase gradient of the distance direction according to the time sequence interferogram.
In an optional embodiment, the superimposing the calculated deformation phase gradient and performing a unwrapping operation on the superimposed deformation phase gradient to obtain an unwrapped absolute phase gradient, including:
respectively superposing the deformation phase gradient in the azimuth direction and the deformation phase gradient in the distance direction to obtain the superposed azimuth deformation phase gradient and distance deformation phase gradient;
and respectively carrying out unwrapping operation on the superposed azimuth deformation phase gradient and the superposed distance deformation phase gradient to obtain an unwrapped azimuth absolute phase gradient and an unwrapped distance absolute phase gradient.
In an optional embodiment, before inputting the unwrapped absolute phase gradient into a pre-trained local surface deformation recognition model and obtaining the recognized deformation region, the method further includes:
constructing a training data set, and dividing the training data set into a training set, a verification set and a test set;
improving a YOLOv3 neural network to obtain an improved YOLOv3 neural network;
and training the local surface deformation recognition model according to the training set, the verification set, the test set and the improved YOLOv3 neural network.
In an alternative embodiment, the YOLOv3 neural network is modified to obtain a modified YOLOv3 neural network, comprising:
and adding an attention mechanism module, a short module and a Drop-Block module in the original Yolov3 neural network to obtain an improved Yolov3 neural network.
In an optional embodiment, an attention mechanism module, a shortcut module and a Drop-Block module are added to the original YOLOv3 neural network to obtain an improved YOLOv3 neural network, which includes:
embedding an attention mechanism module behind each residual block in the YOLOv3 neural network;
the foreshortcut module is connected with the front and back adjacent network layers of the attention mechanism module;
and embedding the Drop-Block module into three detection layers in the YOLOv3 neural network to obtain an improved YOLOv3 neural network.
In an optional embodiment, further comprising:
the training of the model is performed by the CIoU loss function.
In a second aspect, an embodiment of the present application further provides a phase gradient-based surface deformation detection apparatus, including:
the interference pattern generation module is used for obtaining a time sequence interference pattern by adopting the synthetic aperture radar interference measurement;
the phase gradient calculation module is used for calculating a deformation phase gradient according to the time sequence interferogram;
The phase gradient superposition module is used for superposing the calculated deformation phase gradient and carrying out unwrapping operation on the superposed deformation phase gradient to obtain an unwrapped absolute phase gradient;
and the model identification module is used for inputting the unwrapped absolute phase gradient into a pre-trained local earth surface deformation identification model to obtain an identified deformation area.
In a third aspect, an embodiment of the present application provides a phase gradient-based surface deformation detection apparatus, which includes a processor and a memory storing program instructions, where the processor is configured to execute the phase gradient-based surface deformation detection method provided in the foregoing embodiment when executing the program instructions.
In a fourth aspect, the present application provides a computer-readable medium, on which computer-readable instructions are stored, where the computer-readable instructions are executed by a processor to implement a method for detecting surface deformation based on phase gradients provided in the foregoing embodiments.
The technical scheme provided by the embodiment of the application can have the following beneficial effects:
according to the earth surface deformation detection method based on the phase gradient, provided by the embodiment of the application, the unwrapping of a large number of SAR interferograms is avoided through the phase gradient superposition processing, the data processing efficiency is greatly improved, and meanwhile, the computing resource cost is greatly reduced. The influence of atmospheric effect, unwrapping error and the like on the data processing result in the InSAR technology is avoided. The deformation signal in the gradient superposition result is clear and easy to distinguish, the deformation signal is obviously distinguished from weak background information, and the difficulty in distinguishing the deformation signal is reduced. Furthermore, a target detection algorithm in the deep learning network is applied to the detection of the local deformation information in the InSAR, the large-range rapid detection of the local deformation information can be carried out, the reusability is high, and the processing result is not influenced by professional and technical knowledge of data processing personnel. The accuracy and the reliability of earth surface deformation monitoring are improved, and the detection efficiency of local deformation signals in a large range is improved.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
FIG. 1 is a schematic flow diagram illustrating a method for phase gradient based surface deformation detection in accordance with an exemplary embodiment;
FIG. 2 is a schematic diagram illustrating a method for phase gradient based surface deformation detection in accordance with an exemplary embodiment;
FIG. 3 is a schematic diagram illustrating a superimposed morphed phase gradient in accordance with an exemplary embodiment;
FIG. 4 is a schematic diagram illustrating an improved YOLOv3 neural network in accordance with an exemplary embodiment;
FIG. 5 is a schematic diagram of a phase gradient-based surface deformation detection apparatus according to an exemplary embodiment;
FIG. 6 is a schematic diagram illustrating a phase gradient-based surface deformation detection apparatus in accordance with an exemplary embodiment;
FIG. 7 is a schematic diagram illustrating a computer storage medium in accordance with an exemplary embodiment.
Detailed Description
The following description and the drawings sufficiently illustrate specific embodiments of the invention to enable those skilled in the art to practice them.
It should be understood that the described embodiments are only some embodiments of the invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of systems and methods consistent with certain aspects of the invention, as detailed in the appended claims.
With the accumulation of synthetic aperture radar data, important observation support is provided for the detection of surface deformation. However, how to efficiently and quickly process massive SAR image data and acquire local deformation of the earth surface in a wide area still remains a challenge. The method comprises the steps of calculating deformation phase gradients in an interferogram time sequence, and then superposing gradient signals in a time domain to achieve the purpose of enhancing the characteristics of surface deformation signals and separating the surface deformation signals from background noise information. The method avoids unwrapping errors caused by unwrapping a large number of interferograms, greatly inhibits random noise signals while enhancing deformation signals, greatly improves the processing efficiency of SAR data, and greatly reduces errors and required computing resources in data processing, so that the method can be effectively applied to identification of deformation areas in a million square kilometer range. Aiming at deformation information with obvious characteristics and easy distinction, the method improves a classical target detection algorithm YOLOv3 so as to be suitable for efficient automatic detection of a deformation signal target obtained by superposing phase gradients, forms a set of complete technical processes of extraction and automatic detection of local deformation information of the earth surface, and can further provide an effective target area for dynamic monitoring and control of landslide disasters if an active landslide gathering area which is deformed in a large range can be effectively identified.
The method for monitoring surface deformation based on a neural network provided by the embodiment of the present application will be described in detail below with reference to the accompanying drawings. Referring to fig. 1, the method specifically includes the following steps.
S101, obtaining a time series interferogram by adopting synthetic aperture radar interferometry.
The InSAR technology has the outstanding advantages of all-weather, high precision, high efficiency and the like in large-area topographic mapping, and has important application value in the aspect of rapid topographic mapping and updating of topographic maps. In one embodiment, an InSAR technique is used to obtain a time series interferogram of a region under test.
S102, deformation phase gradient is calculated according to the time sequence interferogram.
In an alternative embodiment, the azimuthal deformation phase gradient and the range-wise deformation phase gradient are calculated from the time-series interferogram.
Specifically, the deformation phase gradient of the deformation signal in the interferogram is calculated by utilizing the continuity of the deformation signal in the time and space directions. After the interference measurement obtains the time sequence of the interference pattern, the phase gradient of the azimuth direction and the distance direction is respectively calculated by calculating the phase difference of adjacent pixels in the winding interference pattern. The azimuth direction represents the flight direction of the satellite, the south-north direction, the distance direction represents the direction perpendicular to the azimuth direction, and the east-west direction.
Alternatively, the phase gradient calculation can be performed by using the difference between two pixels separated by a plurality of pixels, or by calculating the gradient in other directions than the azimuth direction and the range direction instead of the adjacent pixel difference method, and then the same deformation signal enhancement effect can be achieved by superimposing the phase gradients of a plurality of interferograms.
S103, the calculated deformation phase gradient is superposed, and the superposed deformation phase gradient is subjected to unwrapping operation to obtain an unwrapped absolute phase gradient.
In an optional embodiment, the azimuthal deformation phase gradient and the range deformation phase gradient are respectively superimposed to obtain a superimposed azimuthal deformation phase gradient and a superimposed range deformation phase gradient.
Specifically, winding phase gradients in the azimuth direction and the distance direction are respectively accumulated and repeatedly wound to [ -pi, pi ], so that the accumulated phase gradients are located in an [ -pi, pi ] interval. The superimposed azimuthal deformation phase gradient and range deformation phase gradient can be calculated respectively according to the following formulas:
wherein,
showing the superimposed phase gradient in the azimuth direction,
the superimposed phase gradient is shown in distance up and Wrap shows the winding operation.
Further, the superposed azimuth deformation phase gradient and range deformation phase gradient are subjected to unwrapping operation respectively to obtain an azimuth absolute phase gradient and a range absolute phase gradient after unwrapping.
In a possible implementation manner, the unwrapping algorithm in the prior art is adopted to perform unwrapping operation on the superposed azimuth deformation phase gradient and range deformation phase gradient respectively to obtain an unwrapped azimuth absolute phase gradient and range absolute phase gradient. Only one unwrapping operation is needed to be carried out on the superposed phases, and unwrapping operation on each interferogram is not needed, so that the complexity of calculation is greatly reduced.
According to the step, aiming at the problem that the extraction of the effective information of the local deformation is difficult in the application of the InSAR method in a wide area range, a method of using a superposed deformation phase gradient is provided, the deformation phase gradient is calculated in an interferogram sequence, and then the superposition of a gradient signal is carried out in a time domain, so that the aim of enhancing the movable deformation signal and separating the movable deformation signal from the background noise information is fulfilled.
In one exemplary scenario, since the gradient of the differential interference phase depends on the topography and deformation of the earth's surface and the difference in atmosphere at the two images, the phase gradient of deformation shows a small scale variation of the deformation gradient, averaging or superimposing the phase gradient map is very useful for reducing the noise level of the data and eliminating atmospheric errors. And the phase difference caused by the atmospheric state is very small between two adjacent pixels, especially after the phase gradient map is superimposed, the time-irrelevant atmospheric signal is further cancelled out, and the deformation gradient of the deformation signal which continuously occurs is continuously enhanced. Thus, the effect of atmospheric turbulence on the superimposed deformation gradient map can be neglected.
Assuming that the earth surface deformation appears as an increase in phase, the phase gradient gradually changes from positive (negative) to negative (positive) along the direction of calculating the phase gradient from the non-deformation (deformation) region to the deformation (non-deformation) region, visually appearing as a gradient characteristic pattern of red-blue coupling. With the increase of the number of the superimposed phase gradient maps, the gradient characteristics can be continuously strengthened,
as shown in fig. 3, (a) shows a result of superimposing a phase gradient in the azimuth direction, and (b) shows a result of superimposing a phase gradient in the range direction. (c) Showing the results of prior art superposition of unwrapped interferograms, of (d) and (e) are magnified views of the rectangular area in (a), showing details of the superimposed phase gradients in azimuth (left) and range (middle), and a comparison of the results of prior art unwrapped phase superposition (right).
Compared to the result obtained by superimposing the unwrapped phases, the local deformation signal (coupled red-blue features) in the phase gradient map can be clearly distinguished from the background noise, the features are more prominent and easier to identify.
Due to the fact that the method avoids unwrapping a large number of interferograms, greatly suppresses noise signals while enhancing deformation signals, and can be effectively applied to the identification of deformation areas in the range of millions of square kilometers. An effective target area can then be provided for analysis and judgment of the deformation accumulation area, such as the active landslide accumulation area.
S104, inputting the unwrapped absolute phase gradient into a pre-trained local earth surface deformation identification model to obtain an identified deformation area.
In one possible implementation, the superimposed phase gradient makes the local deformation signal and the background noise significantly different. The fast localization of the deformation signal becomes a critical issue to be solved. The embodiment of the application is based on a classical target detection algorithm YOLOv3, and the detection task of the deformation signal obtained by phase gradient superposition is correspondingly improved so as to quickly and automatically detect the local deformation region with the phase gradient characteristic.
Specifically, the YOLOv3 neural network is firstly improved, and an attention mechanism module, a short module and a Drop-Block module are added in the original YOLOv3 neural network, so that the improved YOLOv3 neural network is obtained.
In one possible implementation, a Attention mechanism Module (CBAM) is introduced, embedded behind each residual Block in the yollov 3 neural network; meanwhile, in order to prevent gradient dispersion caused by increase of the number of network layers, the shortcut module is connected with the network layers adjacent to the front and the back of the attention mechanism module. As shown in fig. 4a, in the present application, Attention mechanisms of channels and spaces are added to a backbone feature extraction network of YOLOv3, and the capability of the network to extract and screen target key features is improved, which is hereinafter referred to as Attention-YOLOv 3.
Further, to avoid the overfitting phenomenon, the present application applies a regularization method to the Attention-YOLOV 3. Specifically, the Drop-Block module is placed in the three detection layers of the Attention-yollov 3. As shown in FIG. 4b, the Drop-Block module discards neurons in adjacent regions of the feature map together within a certain probability to reduce the fitting ability of the network.
With the above improvement, an improved YOLOv3 neural network is obtained, as shown in fig. 4, (a) represents a residual block with a convolution attention module. (b) The white and dark cells in (1) represent non-activated and activated neurons, respectively, and the black crosses represent randomly lost neurons. (c) A CIoU diagram representing the calculated network loss. In the present embodiment, CIoU loss is used to replace MSE (mean square error) loss in YOLOv 3. The CIoU loss considers the overlapping area between the target detection frames, the distance of the central point and the length-width ratio between the detection frames, and improves the accuracy of target detection.
Further, a training data set is constructed, the training data set is divided into a training set, a verification set and a test set, and the local surface deformation recognition model is trained according to the training set, the verification set, the test set and the improved YOLOv3 neural network.
Specifically, for a data set, local deformation phase gradient anomaly signals with red-blue coupling are manually marked by considering a specific terrain. A total of 712 azimuthal and 581 range gradient images were selected, and the data sets were divided into training, validation and test data sets at an 8:1:1 ratio. A label smoothing method is also applied to smooth the classification labels, and the distance between the labels in the classes is shortened to enhance the generalization capability of the network. The total training duration is 300 rounds. In the first 100 rounds of training, an Adam automatic optimizer is used to find the approximate convergence direction, and in the next 200 rounds, an SGD Momentum optimizer is used to carefully adjust the learning rate to achieve a better training effect. And a k-means clustering method is adopted to divide the width and height of all label targets in the data set into 9 types of anchors so as to reduce the regression loss in the training. Table 1 shows the clustering results:
TABLE 1 Anchor size setting (Unit: Pixel) in training
The MAP after network training is 0.75, and the recall rate and the accuracy rate are 0.87 and 0.95 respectively. Compared with the original YOLOv3 network, the target detection capability is greatly improved.
TABLE.2 deep learning network probe Performance comparison
Through the method, the trained local earth surface deformation identification model is obtained, the unwrapped absolute phase gradient is input into the trained local earth surface deformation identification model, the model outputs four vertex coordinates of the target detection frame of the identified deformation region, and the identified deformation region is obtained.
This application is through effectively combining the new technique of InSAR and the deep learning technique that will provide, can carry out high efficiency to the earth's surface local information in million square kilometers scopes, and is automatic, quick extraction, if to the early recognition of landslide calamity in the national range, compare with traditional InSAR technique, greatly reduced computational resource and human cost, and deformation detection scope is wider.
In order to facilitate understanding of the method for detecting surface deformation based on phase gradient provided in the embodiments of the present application, the following description is made with reference to fig. 2. As shown in fig. 2, the method includes the following steps.
Firstly, obtaining an interferogram sequence through an InSAR technology, calculating deformation phase gradients in the azimuth direction and the distance direction respectively after obtaining a time sequence interferogram, then superposing respectively to obtain a superposed azimuth winding gradient and a superposed distance winding gradient, and then respectively unwrapping the superposed phase gradients to obtain an unwrapped azimuth absolute phase gradient and an unwrapped distance absolute phase gradient.
And further, inputting the unwrapped absolute phase gradient into a trained neural network model, performing local deformation gradient anomaly detection by using a deep learning network, and outputting rectangular coordinate information of a deformation region.
According to the earth surface deformation detection method based on the phase gradient, provided by the embodiment of the application, the unwrapping of a large number of SAR interferograms is avoided through the phase gradient superposition processing, the data processing efficiency is greatly improved, and meanwhile, the computing resource cost is greatly reduced. The influence of atmospheric effect, unwrapping error and the like on the data processing result in the InSAR technology is avoided. The deformation signal in the gradient superposition result is clear and easy to distinguish, the deformation signal is obviously distinguished from weak background information, and the difficulty in distinguishing the deformation signal is reduced. Furthermore, a target detection algorithm in the deep learning network is applied to the detection of the local deformation information in the InSAR, the large-range rapid detection of the local deformation information can be carried out, the reusability is high, and the processing result is not influenced by professional and technical knowledge of data processing personnel. The accuracy and the reliability of earth surface deformation monitoring are improved, and the detection efficiency of local deformation signals in a large range is improved.
An embodiment of the present application further provides a phase gradient-based surface deformation detection apparatus, which is configured to execute the phase gradient-based surface deformation detection method according to the foregoing embodiment, as shown in fig. 5, and the apparatus includes:
An interferogram generating module 501, configured to obtain a time series interferogram by using synthetic aperture radar interferometry;
a phase gradient calculation module 502, configured to calculate a morphometric phase gradient according to the time series interferogram;
a phase gradient superimposing module 503, configured to superimpose the calculated deformation phase gradient, and perform unwrapping operation on the superimposed deformation phase gradient to obtain an unwrapped absolute phase gradient;
and the model identification module 504 is configured to input the unwrapped absolute phase gradient into a pre-trained local earth surface deformation identification model to obtain an identified deformation region.
It should be noted that, when the apparatus for detecting surface deformation based on phase gradient provided in the foregoing embodiment executes the method for detecting surface deformation based on phase gradient, only the division of the above functional modules is taken as an example, and in practical application, the above functions may be allocated to different functional modules according to need, that is, the internal structure of the apparatus may be divided into different functional modules, so as to complete all or part of the above described functions. In addition, the phase gradient-based surface deformation detection device provided by the above embodiment and the phase gradient-based surface deformation detection method embodiment belong to the same concept, and details of implementation processes thereof are referred to in the method embodiment and are not described herein again.
The embodiment of the present application further provides an electronic device corresponding to the method for detecting earth surface deformation based on phase gradient provided in the foregoing embodiment, so as to execute the method for detecting earth surface deformation based on phase gradient.
Please refer to fig. 6, which illustrates a schematic diagram of an electronic device according to some embodiments of the present application. As shown in fig. 6, the electronic apparatus includes: the processor 600, the memory 601, the bus 602 and the communication interface 603, wherein the processor 600, the communication interface 603 and the memory 601 are connected through the bus 602; the memory 601 stores a computer program that can be executed on the processor 600, and the processor 600 executes the computer program to execute the method for detecting surface deformation based on phase gradient provided in any of the foregoing embodiments of the present application.
The Memory 601 may include a high-speed Random Access Memory (RAM) and may further include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. The communication connection between the network element of the system and at least one other network element is realized through at least one communication interface 603 (which may be wired or wireless), and the internet, a wide area network, a local network, a metropolitan area network, and the like can be used.
The electronic device provided by the embodiment of the application and the method for detecting the earth surface deformation based on the phase gradient provided by the embodiment of the application have the same inventive concept and have the same beneficial effects as the method adopted, operated or realized by the electronic device.
Referring to fig. 7, the computer-readable storage medium is an optical disc 700, on which a computer program (i.e., a program product) is stored, and when the computer program is executed by a processor, the computer program performs the method for detecting earth surface deformation based on phase gradient provided in any of the foregoing embodiments.
It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory, or other optical and magnetic storage media, which are not described in detail herein.
The computer-readable storage medium provided by the above embodiment of the present application and the method for detecting surface deformation based on phase gradient provided by the embodiment of the present application have the same inventive concept and have the same beneficial effects as the method adopted, operated or implemented by the application program stored in the computer-readable storage medium.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above examples only show some embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent should be subject to the appended claims.
Claims (10)
1. A method for detecting surface deformation based on phase gradient is characterized by comprising the following steps:
obtaining a time sequence interferogram by adopting synthetic aperture radar interferometry;
calculating a deformation phase gradient according to the time sequence interferogram;
superposing the calculated deformation phase gradient, and performing unwrapping operation on the superposed deformation phase gradient to obtain an unwrapped absolute phase gradient;
And inputting the unwrapped absolute phase gradient into a pre-trained local earth surface deformation identification model to obtain an identified deformation area.
2. The method of claim 1, wherein computing a deformation phase gradient from the time series interferogram comprises:
and calculating the deformation phase gradient in the azimuth direction and the deformation phase gradient in the distance direction according to the time sequence interferogram.
3. The method of claim 2, wherein superimposing the computed morphed phase gradient and performing a unwrapping operation on the superimposed morphed phase gradient to obtain an unwrapped absolute phase gradient comprises:
respectively superposing the deformation phase gradient in the azimuth direction and the deformation phase gradient in the distance direction to obtain the superposed azimuth deformation phase gradient and distance deformation phase gradient;
and respectively carrying out unwrapping operation on the superposed azimuth deformation phase gradient and the superposed distance deformation phase gradient to obtain an unwrapped azimuth absolute phase gradient and an unwrapped distance absolute phase gradient.
4. The method of claim 1, wherein before inputting the unwrapped absolute phase gradient into a pre-trained local surface deformation recognition model to obtain the recognized deformation region, the method further comprises:
Constructing a training data set, and dividing the training data set into a training set, a verification set and a test set;
improving a YOLOv3 neural network to obtain an improved YOLOv3 neural network;
and training the local surface deformation recognition model according to the training set, the verification set, the test set and the improved YOLOv3 neural network.
5. The method of claim 4, wherein the improving the YOLOv3 neural network to obtain an improved YOLOv3 neural network comprises:
and adding an attention mechanism module, a short module and a Drop-Block module in the original Yolov3 neural network to obtain an improved Yolov3 neural network.
6. The method of claim 5, wherein an attention mechanism module, a short module and a Drop-Block module are added to the original YOLOv3 neural network to obtain an improved YOLOv3 neural network, and the method comprises the following steps:
embedding the attention mechanism module behind each residual block in the YOLOv3 neural network;
connecting the front and back adjacent network layers of the attention mechanism module through a shortcut module;
and embedding Drop-Block modules into three detection layers in the YOLOv3 neural network to obtain an improved YOLOv3 neural network.
7. The method of claim 4, further comprising:
training of the model is performed by a CIoU loss function.
8. A surface deformation detection device based on phase gradient is characterized by comprising:
the interference pattern generation module is used for obtaining a time sequence interference pattern by adopting the synthetic aperture radar interference measurement;
the phase gradient calculation module is used for calculating a deformation phase gradient according to the time sequence interferogram;
the phase gradient superposition module is used for superposing the calculated deformation phase gradient and carrying out unwrapping operation on the superposed deformation phase gradient to obtain an unwrapped absolute phase gradient;
and the model identification module is used for inputting the unwrapped absolute phase gradient into a pre-trained local earth surface deformation identification model to obtain an identified deformation area.
9. A phase gradient-based surface deformation detection apparatus comprising a processor and a memory storing program instructions, the processor being configured to perform the phase gradient-based surface deformation detection method according to any one of claims 1 to 7 when executing the program instructions.
10. A computer readable medium having computer readable instructions stored thereon which are executed by a processor to implement a method of detecting earth deformation based on phase gradients as claimed in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210092743.0A CN114563785A (en) | 2022-01-26 | 2022-01-26 | Earth surface deformation detection method, device, equipment and medium based on phase gradient |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210092743.0A CN114563785A (en) | 2022-01-26 | 2022-01-26 | Earth surface deformation detection method, device, equipment and medium based on phase gradient |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114563785A true CN114563785A (en) | 2022-05-31 |
Family
ID=81713208
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210092743.0A Pending CN114563785A (en) | 2022-01-26 | 2022-01-26 | Earth surface deformation detection method, device, equipment and medium based on phase gradient |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114563785A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116148853A (en) * | 2023-02-20 | 2023-05-23 | 中国矿业大学 | Mining area subsidence interference phase filtering method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106772342A (en) * | 2017-01-11 | 2017-05-31 | 西南石油大学 | A kind of Timing Difference radar interference method suitable for big gradient surface subsidence monitoring |
| CN109100720A (en) * | 2018-09-14 | 2018-12-28 | 长安大学 | A kind of InSAR Ground Deformation monitoring method |
| CN110174673A (en) * | 2019-06-24 | 2019-08-27 | 首都师范大学 | A method of it is superimposed using timing relay interference pattern and efficiently weakens atmospheric phase influence |
-
2022
- 2022-01-26 CN CN202210092743.0A patent/CN114563785A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106772342A (en) * | 2017-01-11 | 2017-05-31 | 西南石油大学 | A kind of Timing Difference radar interference method suitable for big gradient surface subsidence monitoring |
| CN109100720A (en) * | 2018-09-14 | 2018-12-28 | 长安大学 | A kind of InSAR Ground Deformation monitoring method |
| CN110174673A (en) * | 2019-06-24 | 2019-08-27 | 首都师范大学 | A method of it is superimposed using timing relay interference pattern and efficiently weakens atmospheric phase influence |
Non-Patent Citations (4)
| Title |
|---|
| XIAOHUA XU 等: "Surface deformation associated with fractures near the 2019 Ridgecrest earthquake sequence", 《SCIENCE》, 30 October 2020 (2020-10-30), pages 605 - 608 * |
| ZHIPENG WU 等: "Deep Learning for the Detection and Phase Unwrapping of Mining-Induced Deformation in Large-Scale Interferograms", 《IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING》, vol. 60, 21 October 2021 (2021-10-21), pages 1 - 18, XP011900650, DOI: 10.1109/TGRS.2021.3121907 * |
| 王生霄 等: "嵌入CBAM结构的改进YOLOV3超宽带雷达生命信号检测算法", 《国外电子测量技术》, vol. 39, no. 03, 15 March 2020 (2020-03-15), pages 1 - 6 * |
| 钱方明 等: "InSAR双轨差分法原理及应用", 《吉林大学学报(工学版)》, vol. 43, no. 1, 15 March 2013 (2013-03-15), pages 92 - 95 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116148853A (en) * | 2023-02-20 | 2023-05-23 | 中国矿业大学 | Mining area subsidence interference phase filtering method and device |
| CN116148853B (en) * | 2023-02-20 | 2024-02-20 | 中国矿业大学 | Mining area subsidence interference phase filtering method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Anantrasirichai et al. | A deep learning approach to detecting volcano deformation from satellite imagery using synthetic datasets | |
| CN112949549B (en) | Super-resolution-based change detection method for multi-resolution remote sensing image | |
| CN111723732B (en) | Optical remote sensing image change detection method, storage medium and computing equipment | |
| Soliman et al. | Effects of vertical accuracy of digital elevation model (DEM) data on automatic lineaments extraction from shaded DEM | |
| CN115236655B (en) | Landslide identification method, system, equipment and medium based on fully-polarized SAR | |
| US8548248B2 (en) | Correlated land change system and method | |
| CN109815997A (en) | The method and relevant apparatus of identification vehicle damage based on deep learning | |
| Li et al. | Road segmentation based on hybrid convolutional network for high-resolution visible remote sensing image | |
| CN111274918B (en) | River dryout and cutoff monitoring method and device based on multi-source remote sensing image | |
| CN101650439B (en) | Method for detecting change of remote sensing image based on difference edge and joint probability consistency | |
| CN110569797B (en) | Method, system and storage medium for detecting mountain fire of geostationary orbit satellite image | |
| Kouchi et al. | Characteristics of tsunami-affected areas in moderate-resolution satellite images | |
| CN115984273B (en) | Road disease detection method, device, computer equipment and readable storage medium | |
| Hordiiuk et al. | Neural network and local laplace filter methods applied to very high resolution remote sensing imagery in urban damage detection | |
| CN114563785A (en) | Earth surface deformation detection method, device, equipment and medium based on phase gradient | |
| Basso et al. | Combining morphological filtering, anisotropic diffusion and block-based data replication for automatically detecting and recovering unscanned gaps in remote sensing images | |
| Carbonneau et al. | Global mapping of river sediment bars | |
| CN117557584A (en) | Water body extraction method and device, electronic equipment and storage medium | |
| CN105631849A (en) | Polygon object change detection method and device | |
| CN110728311A (en) | Image processing method, device and storage medium | |
| CN103903258B (en) | Method for detecting change of remote sensing image based on order statistic spectral clustering | |
| Moughal et al. | An automatic unsupervised method based on context-sensitive spectral angle mapper for change detection of remote sensing images | |
| Zhuang et al. | Filtering approach based on voter model and spatial-contextual information to the binary change map in SAR images | |
| Esch et al. | Monitoring of global urbanization-time series analyses for mega cities based on optical and SAR data | |
| Salmabai et al. | Areal fluctuations monitoring of Al-Azim/Al-Havizeh wetland during the 1986–2017 period, using time-series Landsat data |
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 |