InSAR (Interferometric Synthetic Aperture Radar) co-seismic deformation data set construction method and device for deep learning

Abstract

The embodiment of the invention relates to an InSAR co-seismic deformation data set construction method and device for deep learning. The method comprises the steps that crawling seismic time and crawling seismic information are acquired in response to a crawling request; acquiring earthquake event information matched with the crawling earthquake information in the crawling earthquake time from an earthquake information website, and storing the earthquake event information to an earthquake directory data structure; crawling a frame identification file from an interferogram website, matching epicenter coordinates corresponding to each seismic event in the seismic directory data structure with the frame identification file, and obtaining a frame identification corresponding to each seismic event; and downloading InSAR co-seismic interferogram data from the interferogram website based on the frame identifier corresponding to each seismic event, and processing the InSAR co-seismic interferogram data to obtain an InSAR co-seismic deformation data set. Therefore, the scale and diversity of the data set can be effectively expanded, and high-quality and diversified data support is provided for training of a deep learning model.

Description

Technical Field

[0001] The present disclosure relates to the technical field of data processing, and particularly to a method and apparatus for constructing an InSAR coseismic deformation dataset for deep learning. Background Art

[0002] As a high-precision ground deformation monitoring method, InSAR (Interferometric Synthetic Aperture Radar) technology is widely used in the monitoring and assessment of natural disasters such as earthquakes and volcanoes. Especially after an earthquake, InSAR can provide accurate coseismic deformation images. By monitoring the ground deformation, the earthquake-affected area can be evaluated, and important data support can be provided for post-disaster assessment and recovery work. The advantage of InSAR technology lies in its high spatial resolution and full-coverage ability, making it an important tool in earthquake research. Although InSAR technology has significant advantages in earthquake monitoring and deformation analysis, how to efficiently utilize the massive InSAR dataset remains a major challenge in existing InSAR data processing. Especially in deep learning applications, the scale and quality of the InSAR dataset become the key factors restricting its application potential in deep learning models.

[0003] Traditional methods for constructing coseismic deformation datasets from massive InSAR datasets often rely on manual feature extraction and manual annotation, resulting in relatively limited dataset scale and diversity. These methods usually require a large amount of manual intervention, and the extraction of deformation features depends on experience and domain knowledge. Therefore, with the rise of deep learning technology, in order to meet the requirements of deep learning for large datasets, researchers hope to generate coseismic InSAR datasets through automated means for training deep learning models. The deficiencies of existing datasets make it difficult for deep learning models to fully learn complex deformation patterns, affecting the performance and adaptability of the models. Summary of the Invention

[0004] To solve the above technical problems or at least partially solve the above technical problems, the present disclosure provides a method and apparatus for constructing an InSAR coseismic deformation dataset for deep learning.

[0005] An InSAR co-seismic deformation dataset construction method for deep learning provided by an embodiment of the present disclosure includes: in response to a crawling request, obtaining a crawling earthquake time and crawling earthquake information; obtaining earthquake event information that matches the crawling earthquake information within the crawling earthquake time from an earthquake information website based on a preset crawling technique, and storing the earthquake event information in a preset earthquake catalog data structure; crawling a frame identification file from an interferogram website, and matching the epicenter coordinates corresponding to each earthquake event in the earthquake catalog data structure with the frame identification file to obtain the frame identification corresponding to each earthquake event; downloading InSAR co-seismic interferogram data from the interferogram website based on the frame identification corresponding to each earthquake event, and processing the InSAR co-seismic interferogram data to obtain an InSAR co-seismic deformation dataset.

[0006] Optionally, the method further includes: receiving a start time and an end time input by a user to obtain the crawling earthquake time; receiving a magnitude range, a geographical range, and a focal depth input by the user to obtain the crawling earthquake information; generating the crawling request based on the crawling earthquake time and the crawling earthquake information.

[0007] Optionally, the method further includes: obtaining a data update frequency and data change information of the earthquake information website; adjusting a crawling interval of the earthquake information website based on the data update frequency and the data change information.

[0008] Optionally, the method further includes: performing a screening process on the earthquake event information in the earthquake catalog data structure according to a preset time range and / or space range.

[0009] Optionally, during the process of downloading the InSAR co-seismic interferogram data, the method further includes: determining whether each InSAR co-seismic interferogram is in a storage database; if it is in the storage database, then not performing a download process on the InSAR co-seismic interferogram; when any one of the InSAR co-seismic interferograms fails to be downloaded, performing a re-download process according to a preset number of downloads.

[0010] Optionally, the processing the InSAR co-seismic interferogram data to obtain an InSAR co-seismic deformation dataset includes: screening the InSAR co-seismic interferogram data according to a preset image quality condition to obtain target InSAR co-seismic interferogram data; cropping each target InSAR co-seismic interferogram in the target InSAR co-seismic interferogram data according to a preset cropped image size to obtain to-be-processed InSAR co-seismic interferogram data; performing resampling processing and normalization processing on each of the to-be-processed InSAR co-seismic interferograms to obtain the InSAR co-seismic deformation dataset.

[0011] Optionally, the method further includes: performing image enhancement processing on each image in the InSAR coseismic deformation dataset; wherein, the image enhancement processing includes one or more of image rotation, image flipping, and image scaling and translation.

[0012] An embodiment of the present disclosure also provides an InSAR coseismic deformation dataset construction device for deep learning, including: a response acquisition module, configured to acquire a crawl earthquake time and crawl earthquake information in response to a crawl request; an acquisition and storage module, configured to obtain earthquake event information matching the crawl earthquake information during the crawl earthquake time from an earthquake information website based on a preset crawl technique, and store the earthquake event information into a preset earthquake catalog data structure; a crawl matching module, configured to crawl a frame identification file from an interferogram website, and match the epicenter coordinates corresponding to each earthquake event in the earthquake catalog data structure with the frame identification file to obtain the frame identification corresponding to each earthquake event; a download module, configured to download InSAR coseismic interferogram data from the interferogram website based on the frame identification corresponding to each earthquake event; and an image processing module, configured to process the InSAR coseismic interferogram data to obtain an InSAR coseismic deformation dataset.

[0013] An embodiment of the present disclosure also provides an electronic device, where the electronic device includes: a processor; a memory for storing executable instructions executable by the processor; the processor is configured to read the executable instructions from the memory and execute the instructions to implement the method for constructing an InSAR coseismic deformation dataset for deep learning provided by the embodiment of the present disclosure.

[0014] An embodiment of the present disclosure also provides a computer-readable storage medium, where the storage medium stores a computer program, and the computer program is used to execute the method for constructing an InSAR coseismic deformation dataset for deep learning provided by the embodiment of the present disclosure.

[0015] An embodiment of the present disclosure also provides a computer program product, including a computer program, where the computer program, when executed by a processor, implements the method for constructing an InSAR coseismic deformation dataset for deep learning described in the foregoing aspect.

[0016] The technical solutions provided by the embodiments of the present disclosure have the following advantages compared with the prior art: The InSAR co-seismic deformation dataset construction solution for deep learning provided by the embodiments of the present disclosure includes: in response to a crawling request, obtaining the crawling earthquake time and crawling earthquake information; based on a preset crawling technology, obtaining earthquake event information that matches the crawling earthquake information within the crawling earthquake time from an earthquake information website, and storing the earthquake event information in a preset earthquake catalog data structure; crawling a frame identification file from an interferogram website, and matching the epicenter coordinates corresponding to each earthquake event in the earthquake catalog data structure with the frame identification file to obtain the frame identification corresponding to each earthquake event; downloading InSAR co-seismic interferogram data from the interferogram website based on the frame identification corresponding to each earthquake event, and processing the InSAR co-seismic interferogram data to obtain an InSAR co-seismic deformation dataset. Thus, it is possible to efficiently capture and process earthquake data, avoid the cumbersome process of manual screening and downloading, significantly improve the speed and accuracy of data acquisition, and thus effectively expand the scale and diversity of the dataset, providing high-quality and diverse data support for the training of deep learning models. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In combination with the accompanying drawings and with reference to the following specific embodiments, the above and other features, advantages and aspects of the various embodiments of the present disclosure will become more apparent. Throughout the drawings, the same or similar reference signs denote the same or similar elements. It should be understood that the drawings are schematic and the original components and elements are not necessarily drawn to scale.

[0018] Figure 1 It is a schematic flowchart of a method for constructing an InSAR co-seismic deformation dataset for deep learning provided by an embodiment of the present disclosure;

[0019] Figure 2 It is a schematic flowchart of another method for constructing an InSAR co-seismic deformation dataset for deep learning provided by an embodiment of the present disclosure;

[0020] Figure 3 It is an example diagram of a method for constructing an InSAR co-seismic deformation dataset for deep learning provided by an embodiment of the present disclosure;

[0021] Figure 4 It is a schematic structural diagram of an InSAR co-seismic deformation dataset construction device for deep learning provided by an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Instead, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the accompanying drawings and embodiments of the present disclosure are only for illustrative purposes and are not used to limit the protection scope of the present disclosure.

[0023] It should be understood that the various steps recited in the method embodiments of the present disclosure can be executed in a different order and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this regard.

[0024] As used herein, the term "including" and its variations are open-ended, that is, "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0025] It should be noted that the concepts such as "first", "second", etc. mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order of the functions performed by these devices, modules or units or the interdependence relationship.

[0026] It should be noted that the modifications of "one" and "plural" mentioned in the present disclosure are illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise clearly stated in the context, it should be understood as "one or more".

[0027] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0028] Based on the foregoing description of the background art, the download and processing of InSAR co-seismic deformation data often rely on manual intervention. Especially when obtaining data from public databases, it is still necessary to manually select the time, location, and relevant data, resulting in cumbersome operations, low efficiency, and small scale. For example, the current processing of InSAR data usually requires manually obtaining relevant earthquake catalogs and downloading corresponding interferogram data. Operations such as screening, cropping, and resampling may be involved in the processing. This mode that relies on manual operations not only increases the workload of data processing but also is prone to human errors, affecting the quality and consistency of the dataset. In addition, the existing technology lacks an efficient systematic solution for the automated processing of large-scale datasets, resulting in inefficient and repetitive data generation and enhancement processes. The lack of automated data scraping and processing means greatly reduces the efficiency and scalability of InSAR data in large-scale and real-time applications.

[0029] In view of the above problems, the present disclosure proposes a method for constructing an InSAR co-seismic deformation dataset for deep learning. By automatically crawling earthquake information on earthquake information websites such as the GCMT website and downloading corresponding co-seismic interferogram data from interferogram websites such as the LiCSAR website according to this earthquake information, and combining processing steps such as data screening, cropping, resampling, and normalization, a diverse and large-scale InSAR dataset is generated. And by combining data augmentation techniques (such as translation, scaling, rotation, etc.), the scale and diversity of the dataset are further expanded to support the efficient training of deep learning models for complex co-seismic deformations, overcoming the problems in the prior art such as small dataset scale, insufficient sample diversity, low data processing efficiency, and the deep learning model relying on high-quality data; the embodiments of the present disclosure can provide rich training data for deep learning models, improve the generalization ability and accuracy of the models, and provide effective support for the practical application of InSAR technology in earthquake monitoring.

[0030] Figure 1 FIG. is a schematic flowchart of a method for constructing an InSAR co-seismic deformation dataset for deep learning provided by an embodiment of the present disclosure. This method can be executed by an InSAR co-seismic deformation dataset construction device for deep learning, where the device can be implemented by software and / or hardware and is generally integrated in an electronic device. As Figure 1 shown, the method includes:

[0031] Step 101: In response to a crawling request, obtain the crawling earthquake time and crawling earthquake information.

[0032] Step 102: Based on a preset crawling technique, obtain earthquake event information that matches the crawling earthquake information within the crawling earthquake time from the earthquake information website, and store the earthquake event information in a preset earthquake catalog data structure.

[0033] In an embodiment of the present disclosure, it is possible to receive the start time and end time input by the user, obtain the earthquake crawling time, and receive the magnitude range, geographical range, and focal depth input by the user, obtain the earthquake crawling information, and generate a crawling request based on the earthquake crawling time and the earthquake crawling information.

[0034] In an embodiment of the present disclosure, the start time and end time provided by the user are received as the earthquake crawling time, the magnitude range, geographical range (latitude and longitude), focal depth, etc. are received as the earthquake crawling information, and the earthquake information website such as the GCMT website is automatically accessed according to the earthquake crawling time and the earthquake crawling information.

[0035] In an embodiment of the present disclosure, a crawling technique is preset to obtain earthquake event information matching the earthquake crawling information within the earthquake crawling time from the earthquake information website, such as earthquake event information including the earthquake occurrence time, epicenter coordinates, magnitude, focal depth, etc., and save the earthquake event information (such as the earthquake occurrence time, epicenter coordinates (latitude, longitude)) to the earthquake catalog data structure; wherein, the earthquake catalog data structure at least includes the earthquake occurrence time and epicenter coordinates (latitude, longitude) corresponding to each earthquake event.

[0036] In some embodiments, it is also possible to obtain the data update frequency and data change information of the earthquake information website, and adjust the crawling interval of the earthquake information website based on the data update frequency and data change information. For example, the higher the data update frequency and the more the data change information, the shorter the crawling interval of the earthquake information website, so as to set a reasonable crawling interval according to the data update frequency and data change situation of the earthquake information website, and further ensure the integrity and accuracy of the data.

[0037] In some embodiments, it is also possible to perform screening processing on the earthquake event information in the earthquake catalog data structure according to a preset time range and / or space range. That is to say, after the earthquake data is crawled, the earthquake data will be preliminarily screened to remove the earthquake data that does not meet the requirements in terms of time or space range, so as to ensure that the subsequent InSAR data download request highly matches the actual demand.

[0038] Step 103: Crawl the frame identification file from the interferogram website, and match the epicenter coordinates corresponding to each earthquake event in the earthquake catalog data structure with the frame identification file to obtain the frame identification corresponding to each earthquake event.

[0039] Step 104: Download the InSAR coseismic interferogram data from the interferogram website based on the frame identification corresponding to each earthquake event, and process the InSAR coseismic interferogram data to obtain the InSAR coseismic deformation data set.

[0040] In the embodiments of the present disclosure, after crawling the earthquake information website, the geographical area information (latitude and longitude) corresponding to each FrameID in the interferogram website such as the LiCSAR website will be automatically crawled to ensure accurate positioning to the correct area when downloading co-seismic interferograms from the interferogram website subsequently; the geographical area information corresponding to each FrameID will be stored and matched with the information in the earthquake catalog data structure to provide accurate positioning data for subsequent co-seismic interferogram data download.

[0041] Specifically, after the earthquake event information and the frame identification file are crawled, according to the epicenter coordinates of each earthquake event, the frame identification corresponding to the earthquake is obtained from the frame identification file, and then a download request is automatically sent to the interferogram website through the interferogram download control area; since the interferogram website provides all InSAR interferograms, according to information such as the earthquake occurrence time, the corresponding co-seismic interferogram data can be accurately matched.

[0042] In some embodiments, during the process of downloading InSAR co-seismic interferogram data, it can also be determined whether each InSAR co-seismic interferogram is in the storage database. If it is in the storage database, the InSAR co-seismic interferogram will not be downloaded. When any InSAR co-seismic interferogram fails to download, it will be redownloaded according to the preset number of downloads.

[0043] Specifically, the downloaded co-seismic interferograms are stored in the local data storage database, and the source of the data (information such as which frame identification and the interference time) is marked to ensure the traceability of the data. This process will verify whether the data already exists. If it already exists, the download will be skipped and the interferogram of the next earthquake will be downloaded continuously. If there is a download failure or the data is incomplete, an automatic retry will be performed to ensure that the data requested each time is complete.

[0044] In the embodiments of the present disclosure, the InSAR co-seismic interferogram data is processed to obtain an InSAR co-seismic deformation dataset, including: screening the InSAR co-seismic interferogram data according to the preset image quality conditions to obtain the target InSAR co-seismic interferogram data; cropping each target InSAR co-seismic interferogram in the target InSAR co-seismic interferogram data according to the preset cropped image size to obtain the to-be-processed InSAR co-seismic interferogram data; performing resampling processing and normalization processing on each to-be-processed InSAR co-seismic interferogram to obtain the InSAR co-seismic deformation dataset.

[0045] Specifically, the acquired InSAR coseismic interferogram data usually contain image files of large sizes, and the interferograms of different earthquake events may have problems such as different resolutions, image sizes, noises, etc. Therefore, it is necessary to quickly screen and eliminate invalid or poor-quality images, and divide the large-size InSAR interferograms into small images, aiming to improve the calculation efficiency and reduce the memory consumption. Interferograms from different sources may have different resolutions. To unify the data format and facilitate subsequent analysis, resampling operations are performed on the cropped images. To eliminate the differences between different images, the system normalizes all images so that the data can be compared under the same standard, ensuring the unity of the data.

[0046] In the embodiment of the present disclosure, image enhancement processing is performed on each image in the InSAR coseismic deformation dataset; wherein, the image enhancement processing includes one or more of image rotation, image flipping, and image scaling and translation.

[0047] Specifically, when generating the processed dataset, in order to increase the scale and diversity of the dataset, the present disclosure adopts data augmentation technology; data augmentation is a technology that generates new data by transforming existing data, which can not only expand the size of the dataset, but also enable the model to adapt to more types of coseismic deformation features; specific augmentation methods include one or more of image rotation, image flipping, and image scaling and translation.

[0048] The InSAR coseismic deformation dataset construction scheme for deep learning provided by the embodiment of the present disclosure, in response to a crawling request, acquires the crawled earthquake time and crawled earthquake information; based on a preset crawling technology, obtains earthquake event information matching the crawled earthquake information during the crawled earthquake time from an earthquake information website, and stores the earthquake event information into a preset earthquake catalog data structure; crawls a frame identification file from an interferogram website, and matches the epicenter coordinates corresponding to each earthquake event in the earthquake catalog data structure with the frame identification file to obtain the frame identification corresponding to each earthquake event; downloads InSAR coseismic interferogram data from the interferogram website based on the frame identification corresponding to each earthquake event, and processes the InSAR coseismic interferogram data to obtain an InSAR coseismic deformation dataset. Thus, earthquake data can be efficiently grabbed and processed, avoiding the cumbersome process of manual screening and downloading, significantly improving the speed and accuracy of data acquisition, and thus being able to effectively expand the scale and diversity of the dataset, providing high-quality and diverse data support for the training of deep learning models.

[0049] Specifically, Figure 2Schematic flowchart of another method for constructing an InSAR co-seismic deformation dataset for deep learning provided by an embodiment of the present disclosure. Based on the above embodiment, the method for constructing an InSAR co-seismic deformation dataset for deep learning is further optimized. As Figure 2 shown, the method includes:

[0050] Step 201: Receive the start time and end time input by the user to obtain the crawling earthquake time, receive the magnitude range, geographical range, and focal depth input by the user to obtain the crawling earthquake information, and generate a crawling request based on the crawling earthquake time and the crawling earthquake information.

[0051] Step 202: In response to the crawling request, obtain the crawling earthquake time and the crawling earthquake information, obtain earthquake event information that matches the crawling earthquake information within the crawling earthquake time from an earthquake information website based on a preset crawling technique, and store the earthquake event information in a preset earthquake catalog data structure.

[0052] Step 203: Obtain the data update frequency and data change information of the earthquake information website, adjust the crawling interval of the earthquake information website based on the data update frequency and the data change information, and perform screening processing on the earthquake event information in the earthquake catalog data structure according to a preset time range and / or space range.

[0053] Step 204: Crawl the frame identification file from the interferogram website, and match the epicenter coordinates corresponding to each earthquake event in the earthquake catalog data structure with the frame identification file to obtain the frame identification corresponding to each earthquake event.

[0054] Step 205: Download InSAR co-seismic interferogram data from the interferogram website based on the frame identification corresponding to each earthquake event. During the process of downloading the InSAR co-seismic interferogram data, determine whether each InSAR co-seismic interferogram is in the storage database. If it is in the storage database, no download processing is performed on the InSAR co-seismic interferogram. When any InSAR co-seismic interferogram download fails, re-download it according to a preset number of downloads.

[0055] Step 206: Screen the InSAR co-seismic interferogram data according to a preset image quality condition to obtain the target InSAR co-seismic interferogram data, crop each target InSAR co-seismic interferogram in the target InSAR co-seismic interferogram data according to a preset cropped image size to obtain the to-be-processed InSAR co-seismic interferogram data, and perform resampling processing and normalization processing on each to-be-processed InSAR co-seismic interferogram to obtain the InSAR co-seismic deformation dataset.

[0056] Step 207: Perform image enhancement processing on each image in the InSAR coseismic deformation dataset; among them, the image enhancement processing includes one or more of image rotation, image flipping, and image scaling and translation.

[0057] The earthquake catalog crawling, automatic generation of coseismic interferogram dataset, and data enhancement proposed in this disclosure take the GCMT website as an example of the earthquake information website and the LiCSAR website as an example of the interferogram website. As an example, as Figure 3 shown, it includes the following steps:

[0058] (1) Initialize the data scraping and processing module. In order to efficiently and automatically process large-scale InSAR data, this disclosure first designs a data scraping and processing module. This module scrapes the earthquake catalog data in the GCMT website, and then obtains the relevant InSAR coseismic interferograms.

[0059] Specifically, in the embodiment of this disclosure, the earthquake catalog data structure is preset in advance. This earthquake catalog data structure is used to store the earthquake event information scraped from the GCMT website, including the time of earthquake occurrence, epicenter coordinates (latitude, longitude), magnitude, depth, etc.; through this earthquake catalog data structure, earthquake events that meet the conditions can be screened out according to the given time range, supporting users to flexibly select the analysis time period, and finally only the time of earthquake occurrence and epicenter coordinates (latitude, longitude) are saved.

[0060] Specifically, in the embodiment of this disclosure, there is also a data screening area. After scraping data such as earthquake information, the data is preliminarily screened to remove data that does not meet the requirements in terms of time or space range, so as to ensure that the subsequent InSAR data download requests highly match the actual needs.

[0061] Specifically, in the embodiment of this disclosure, there is also an interferogram download control area. This interferogram download control area is mainly used to manage the coseismic interferogram data requests downloaded from the LiCSAR website. Through this interferogram download control area, it is ensured that the data content of each request is accurate and error-free, avoiding invalid or repeated downloads.

[0062] Specifically, Figure 3 It also includes the step: (2) Crawl the GCMT earthquake catalog to realize the automatic process of crawling the earthquake catalog from the GCMT website. First, receive the information provided by the user, such as the start time and end time, magnitude range, geographical range (latitude and longitude), focal depth, etc., and automatically access the GCMT website according to the above information; use web crawler technology to automatically extract the earthquake event information that meets the conditions and store it in the earthquake catalog data structure; in order to ensure the integrity and accuracy of the data, set a reasonable scraping interval according to the update frequency and data change situation of the GCMT website.

[0063] It should be noted that during this process, earthquake event information that meets the requirements is screened, and earthquake events that match the user's requirements are captured and stored; these earthquake events include, but are not limited to, key information such as the time of the earthquake, epicenter coordinates, magnitude, depth, etc., and the earthquake catalog information that meets the requirements (the time of the earthquake, epicenter coordinates (latitude, longitude)) is saved to a document for subsequent downloading and processing of relevant InSAR co-seismic interferogram data.

[0064] Specifically, Figure 3 It also includes the step: (3) Crawl the FrameID file. After crawling the GCMT earthquake catalog, the geographical area information (latitude and longitude) corresponding to each FrameID in the LiCSAR website will be automatically crawled; this process is used to ensure that the correct area can be accurately located when downloading co-seismic interferograms from the LiCSAR website in the future; the geographical area information corresponding to each FrameID will be stored and matched with the information in the earthquake catalog to provide accurate positioning data for subsequent downloading of co-seismic interferogram data.

[0065] Specifically, Figure 3 It also includes the step: (4) Download LiCSAR co-seismic interferogram data. When the earthquake event information and FrameID file crawling are completed, the present disclosure will obtain the FrameID corresponding to the earthquake from the FrameID file according to the epicenter coordinates of each earthquake event, and then send a download request to the LiCSAR website automatically through the interferogram download control area. The LiCSAR website provides all InSAR interferograms of Sentinel-1 satellite data. Therefore, the system will accurately match the corresponding co-seismic interferogram data according to the earthquake occurrence time information.

[0066] Specifically, once the relevant data is obtained, the downloaded co-seismic interferograms are stored in the local data storage module, and the source of the data (information such as which FrameID it comes from and the interference time, etc.) is marked to ensure the traceability of the data; this process will verify whether the data already exists. If it already exists, the download will be skipped and the interferogram of the next earthquake will be downloaded; if the download fails or the data is incomplete, the system will automatically retry to ensure that the data requested each time is complete.

[0067] Specifically, Figure 3It also includes the steps: (5) Interferogram data processing. The acquired InSAR co-seismic interferogram data usually contains image files of relatively large sizes, and interferograms of different earthquake events may have problems such as different resolutions, image sizes, noises, etc. Therefore, after downloading the interferogram data, a series of preprocessing operations are performed on the data, including: Screening. First, through the pop-up selection window, the system quickly screens and eliminates invalid or low-quality images. For example, if an image has severe decoherence or obvious noise, the system will automatically retain the higher-quality image data to ensure the high accuracy of the data for subsequent processing; Cropping. The large-size InSAR interferogram is segmented into small images, aiming to improve the calculation efficiency and reduce memory consumption. By focusing on specific areas related to earthquakes or deformations, this step can effectively reduce the interference of irrelevant areas and ensure the spatial consistency and scale standardization of the data. In addition, this processing helps to improve the operability of the data and provide higher-quality input data for the training of subsequent deep learning models; Resampling. Interferograms from different sources may have different resolutions. To unify the data format and facilitate subsequent analysis, the system performs resampling operations on the cropped images. To conform to the standard resolution for deep learning, the resolutions of all images are resampled to the target pixel size, such as 224×224 pixels; Normalization processing. To eliminate the differences between different images, the system performs normalization processing on all images, enabling the data to be compared under the same standard, ensuring the unity of the data, and improving the stability and convergence speed of the deep learning model. The above data processing operations can ensure the consistency and comparability of the interferogram data in the subsequent training of the deep learning model.

[0068] Specifically, Figure 3 It also includes the steps: (6) Data augmentation. When generating the processed dataset, to increase the scale and diversity of the dataset, the present disclosure adopts data augmentation techniques. Data augmentation is a technique that generates new data by transforming existing data, which can not only expand the size of the dataset but also enable the model to adapt to more types of co-seismic deformation features. Specific augmentation methods include: Image rotation. By randomly rotating the interferogram images, the co-seismic deformations under different perspectives can be simulated, enhancing the model's adaptability to different earthquake scenarios; Image flipping. Randomly flipping the image horizontally or vertically to increase the diversity of the samples; Image scaling and translation. By randomly scaling or translating the image, different scales of earthquake deformations are simulated; Combinatorial transformation. Randomly combining the above multiple transformation forms; Through these simple yet effective image transformations in the data augmentation step, the originally small-scale dataset is significantly expanded, providing more training samples for the deep learning model to learn.

[0069] Specifically, Figure 3It also includes the steps: (7) Dataset generation and output. Through the above data processing and enhancement process, the present disclosure finally generates a diverse and large-scale InSAR coseismic deformation dataset. This dataset contains coseismic deformation images of multiple earthquake events, covering earthquake information of different geographical locations, different epicenter coordinates, and different magnitudes. The generated dataset not only meets the requirements of the deep learning model for a large number of samples but also has sufficient diversity to effectively support the training and validation of the model.

[0070] Thus, by developing an integrated data scraping and processing module, the automatic scraping and screening of the earthquake event catalog on the GCMT website are realized. Combined with the user-specified time range, it ensures the accurate acquisition of the required earthquake information, significantly improves the data scraping efficiency, and reduces manual intervention. At the same time, through the integration with the LiCSAR website, it can automatically match and download the corresponding InSAR coseismic interferogram data according to the earthquake event information, avoiding repeated downloads and saving storage resources, and improving the overall operation efficiency of the system. To meet the requirements of the deep learning model, the system performs multiple steps of preprocessing on the downloaded coseismic interferogram data, including data screening, cropping, resampling, and normalization, ensuring the consistency of the dataset and eliminating the differences between different images. In addition, by combining multiple data enhancement techniques (such as image rotation, flipping, scaling, etc.), the present invention can effectively expand the scale and diversity of the dataset, providing high-quality and diverse data support for the training of the deep learning model.

[0071] Compared with the prior art, the present disclosure efficiently scrapes and processes earthquake data from the GCMT and LiCSAR websites through an automatic crawler and a data processing module, avoiding the cumbersome process of manual screening and downloading, and significantly improving the speed and accuracy of data acquisition. The data screening, cropping, resampling, and normalization processing steps ensure that the generated dataset has high quality and consistency. In addition, by combining data enhancement techniques such as image rotation, flipping, and scaling, the scale and diversity of the dataset are expanded, which helps to improve the robustness and adaptability of the deep learning model, especially the application effect in the coseismic deformation detection task.

[0072] Figure 4 It is a schematic structural diagram of a device for constructing an InSAR coseismic deformation dataset for deep learning provided by an embodiment of the present disclosure. This device can be implemented by software and / or hardware and is generally integrated in an electronic device. As Figure 4 shown, the device includes:

[0073] A response acquisition module 401, configured to acquire the crawled earthquake time and crawled earthquake information in response to a crawling request.

[0074] Obtain a storage module 402, which is used to obtain earthquake event information matching the crawled earthquake information within the crawled earthquake time from an earthquake information website based on a preset crawling technique, and store the earthquake event information into a preset earthquake catalog data structure;

[0075] A crawling and matching module 403, which is used to crawl a frame identification file from an interferogram website, and match the epicenter coordinates corresponding to each earthquake event in the earthquake catalog data structure with the frame identification file to obtain the frame identification corresponding to each earthquake event;

[0076] A download module 404, which is used to download InSAR coseismic interferogram data from the interferogram website based on the frame identification corresponding to each earthquake event;

[0077] An image processing module 405, which is used to process the InSAR coseismic interferogram data to obtain an InSAR coseismic deformation data set.

[0078] Optionally, the device further includes: a receiving module, which is used to receive the start time and end time input by the user to obtain the crawled earthquake time, and receive the magnitude range, geographical range, and focal depth input by the user to obtain the crawled earthquake information; a generating module, which is used to generate the crawling request based on the crawled earthquake time and the crawled earthquake information.

[0079] Optionally, the device further includes: an obtaining and adjusting module, which is used to obtain the data update frequency and data change information of the earthquake information website, and adjust the crawling interval of the earthquake information website based on the data update frequency and the data change information.

[0080] Optionally, the method further includes: a screening and processing module, which is used to screen and process the earthquake event information in the earthquake catalog data structure according to a preset time range and / or space range.

[0081] Optionally, during the process of downloading the InSAR coseismic interferogram data, the device further includes: a judgment and processing module, which is used to judge whether each InSAR coseismic interferogram is in the storage database. If it is in the storage database, the InSAR coseismic interferogram is not downloaded. When any InSAR coseismic interferogram fails to be downloaded, it is redownloaded according to a preset number of download times.

[0082] Optionally, the image processing module is specifically configured to: screen the InSAR coseismic interferogram data according to a preset image quality condition to obtain target InSAR coseismic interferogram data; crop each target InSAR coseismic interferogram in the target InSAR coseismic interferogram data according to a preset cropped image size to obtain the to-be-processed InSAR coseismic interferogram data; perform resampling processing and normalization processing on each to-be-processed InSAR coseismic interferogram to obtain the InSAR coseismic deformation data set.

[0083] Optionally, the image processing module is further specifically configured to: perform image enhancement processing on each image in the InSAR coseismic deformation data set; wherein, the image enhancement processing includes one or more of image rotation, image flipping, and image scaling and translation.

[0084] The InSAR coseismic deformation data set construction device for deep learning provided by the embodiments of the present disclosure can execute the InSAR coseismic deformation data set construction method for deep learning provided by any embodiment of the present disclosure, and has the corresponding functional modules and beneficial effects for executing the method.

[0085] The embodiments of the present disclosure also provide a computer program product, including a computer program / instructions, and when the computer program / instructions are executed by a processor, the InSAR coseismic deformation data set construction method provided by any embodiment of the present disclosure is implemented.

[0086] It should be noted that the computer-readable medium described above in the present disclosure can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of a computer-readable storage medium can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present disclosure, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on a computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0087] In some embodiments, the client and the server can communicate using any currently known or future-developed network protocol such as HTTP (Hyper Text Transfer Protocol), and can be interconnected with digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include local area networks (“LAN”), wide area networks (“WAN”), the Internet (e.g., the Internet), and end-to-end networks (e.g., ad hoc end-to-end networks), as well as any currently known or future-developed network.

[0088] The above computer-readable medium can be included in the above electronic device; or it can exist separately without being assembled into the electronic device.

[0089] The above computer-readable medium carries one or more programs, and when the one or more programs are executed by the electronic device, the electronic device performs the method described in the foregoing embodiments.

[0090] Computer program code for performing the operations of this disclosure may be written in one or more programming languages or combinations thereof. The programming languages include, but are not limited to, object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code may execute entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0091] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a part of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions noted in the blocks may occur in a different order than noted in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, may be implemented by a dedicated hardware-based system that performs the specified functions or operations, or may be implemented by a combination of dedicated hardware and computer instructions.

[0092] The units involved in the embodiments described in this disclosure may be implemented in software or in hardware. Wherein, the name of the unit does not constitute a limitation on the unit itself in some cases.

[0093] The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGA), application specific integrated circuits (ASIC), application specific standard products (ASSP), system on a chip (SOC), complex programmable logic devices (CPLD), and so on.

[0094] In the context of the present disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium would include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0095] According to one or more embodiments of the present disclosure, the present disclosure provides an electronic device, including:

[0096] A processor;

[0097] A memory for storing executable instructions of the processor;

[0098] The processor is configured to read the executable instructions from the memory and execute the instructions to implement any one of the InSAR coseismic deformation dataset construction methods for deep learning provided by the present disclosure.

[0099] According to one or more embodiments of the present disclosure, the present disclosure provides a computer-readable storage medium storing a computer program for executing any one of the InSAR coseismic deformation dataset construction methods for deep learning provided by the present disclosure.

[0100] The above description is only a preferred embodiment of the present disclosure and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosure concept. For example, the technical solutions formed by mutually replacing the above features with other technical features having similar functions disclosed in the present disclosure (but not limited to).

[0101] Moreover, although the operations are depicted in a particular order, this should not be construed as requiring that the operations be performed in the particular order shown or in sequential order. In certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although several specific implementation details are included in the above discussion, these should not be construed as limitations on the scope of the present disclosure. Certain features that are described in the context of separate embodiments may also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment may also be implemented separately or in any suitable sub-combination in multiple embodiments.

[0102] Although the subject matter has been described in language specific to structural features and / or methodological acts, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are merely example forms of implementing the claims.

Claims

1. A method for constructing an InSAR coseismic deformation dataset for deep learning, characterized in that, The method includes: In response to a crawling request, obtaining the crawling earthquake time and the crawling earthquake information; Based on a preset crawling technique, obtaining earthquake event information that matches the crawling earthquake information within the crawling earthquake time from an earthquake information website, and storing the earthquake event information into a preset earthquake catalog data structure; Crawling a frame identification file from an interferogram website, and matching the epicenter coordinates corresponding to each earthquake event in the earthquake catalog data structure with the frame identification file to obtain the frame identification corresponding to each earthquake event; Based on the frame identification corresponding to each earthquake event, downloading InSAR coseismic interferogram data from the interferogram website, and processing the InSAR coseismic interferogram data to obtain an InSAR coseismic deformation dataset.

2. The method according to claim 1, wherein The method further includes: Receiving a start time and an end time input by a user, and obtaining the crawling earthquake time; Receiving a magnitude range, a geographical range, and a focal depth input by the user, and obtaining the crawling earthquake information; Generating the crawling request based on the crawling earthquake time and the crawling earthquake information.

3. The method according to claim 1, wherein The method further includes: Obtaining the data update frequency and data change information of the earthquake information website; Adjusting the crawling interval of the earthquake information website based on the data update frequency and the data change information.

4. The method according to claim 1, characterized in that, The method further includes: Performing a screening process on the earthquake event information in the earthquake catalog data structure according to a preset time range and / or space range.

5. The method according to claim 1, characterized in that, During the process of downloading the InSAR coseismic interferogram data, the method further includes: Determining whether each InSAR coseismic interferogram is in a storage database; If it is in the storage database, then not performing a download process on the InSAR coseismic interferogram; When any of the InSAR coseismic interferograms fails to be downloaded, performing a re-download process according to a preset number of downloads.

6. The method according to claim 1, wherein The processing of the InSAR coseismic interferogram data to obtain an InSAR coseismic deformation dataset includes: Screening the InSAR coseismic interferogram data according to preset image quality conditions to obtain target InSAR coseismic interferogram data; Cropping each target InSAR coseismic interferogram in the target InSAR coseismic interferogram data according to a preset cropped image size to obtain data of the InSAR coseismic interferogram to be processed; Performing resampling processing and normalization processing on each of the data of the InSAR coseismic interferogram to be processed to obtain the InSAR coseismic deformation dataset.

7. The method according to claim 6, wherein The method further includes: Performing image enhancement processing on each image in the InSAR coseismic deformation dataset; wherein, the image enhancement processing includes one or more of image rotation, image flipping, and image scaling and translation.

8. An InSAR co-seismic deformation dataset construction device for deep learning, characterized in that, Includes: A response acquisition module, configured to obtain the crawling earthquake time and the crawling earthquake information in response to a crawling request; An acquisition and storage module, configured to obtain earthquake event information that matches the crawling earthquake information within the crawling earthquake time from an earthquake information website based on a preset crawling technique, and store the earthquake event information into a preset earthquake catalog data structure; A crawling and matching module, configured to crawl a frame identification file from an interferogram website, and match the epicenter coordinates corresponding to each earthquake event in the earthquake catalog data structure with the frame identification file, so as to obtain the frame identification corresponding to each earthquake event; A downloading module, configured to download InSAR coseismic interferogram data from the interferogram website based on the frame identification corresponding to each earthquake event; An image processing module, configured to process the InSAR coseismic interferogram data to obtain an InSAR coseismic deformation data set.

9. An electronic device, characterized in that, The electronic device includes: A processor; A memory for storing executable instructions of the processor; The processor is configured to read the executable instructions from the memory and execute the instructions to implement the method for constructing an InSAR coseismic deformation data set for deep learning according to any one of claims 1-7 above.

10. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, and the computer program is used to execute the method for constructing an InSAR coseismic deformation data set for deep learning according to claims 1-7 above.

Images