Multi-scene analysis method and device for earth surface deformation point cloud data

Abstract

The embodiment of the invention discloses a multi-scene analysis method and device for earth surface deformation point cloud data. The method comprises the following steps: constructing an initial earth surface deformation analysis model, and carrying out segmented training on the initial earth surface deformation analysis model based on a multi-source earth surface deformation point cloud set to obtain a trained earth surface deformation analysis model; then, target analysis data and a target analysis scene corresponding to the received analysis instruction are determined according to an instruction analysis method, and a model coefficient corresponding to the target analysis scene is determined based on a scene configuration library; and finally, determining an analysis report corresponding to the target analysis data and the model coefficient according to the trained surface deformation analysis model. Through the above customized training and targeted adjustment of the constructed earth surface deformation analysis model according to the analysis instruction, the field adaptability of the analysis model is increased, the multi-scene analysis demand of the earth surface deformation point cloud data is met, and the user satisfaction is improved.

Description

Technical Field

[0001] This specification relates to the field of data processing technology in one or more embodiments, and in particular to a multi-scenario analysis method for surface deformation point cloud data. Background Technology

[0002] Ground subsidence, ground fissures, landslides, and other surface deformations are core risk factors in geological disasters and urban construction. Synthetic aperture radar interferometry can generate massive amounts of surface deformation point cloud data, providing high-precision data support to ensure the safety of buildings and people's property. Therefore, a method is needed to analyze large amounts of surface deformation point cloud data in a timely manner. However, existing analysis technologies have shortcomings such as insufficient management and fusion of multi-source data, lack of standardization in professional analysis and large-scale model collaboration, and insufficient ability of general large-scale models to adapt to surface deformation scenarios. These shortcomings make it impossible to meet the rapid analysis requirements of different scenarios at the same time, thus reducing user satisfaction. Summary of the Invention

[0003] This specification provides a method and apparatus for multi-scene analysis of surface deformation point cloud data, the technical solution of which is as follows:

[0004] Firstly, embodiments of this specification provide a multi-scene analysis method for surface deformation point cloud data, the method comprising:

[0005] An initial surface deformation analysis model is constructed, and the initial surface deformation analysis model is trained in segments based on a multi-source surface deformation point cloud to obtain a trained surface deformation analysis model. The multi-source surface deformation point cloud includes a surface deformation point cloud dataset and a labeled dataset.

[0006] The target analysis data and target analysis scenario corresponding to the received analysis command are determined according to the command analysis method, and the model coefficients corresponding to the target analysis scenario are determined based on the scenario configuration library. The model coefficients include spatiotemporal attention weight coefficients, domain knowledge attention coefficients, and table dimension selection coefficients.

[0007] Based on the trained surface deformation analysis model, an analysis report corresponding to the target analysis data and model coefficients is determined. The analysis report includes deformation phenomena, causes of deformation, and deformation risks.

[0008] Secondly, a multi-scene analysis device for surface deformation point cloud data is provided, the device comprising:

[0009] The construction module is used to construct an initial surface deformation analysis model and to perform segmented training on the initial surface deformation analysis model based on a multi-source surface deformation point cloud to obtain a trained surface deformation analysis model. The multi-source surface deformation point cloud includes a surface deformation point cloud dataset and a labeled dataset.

[0010] The determination module is used to determine the target analysis data and target analysis scenario corresponding to the received analysis command according to the command analysis method, and to determine the model coefficients corresponding to the target analysis scenario based on the scenario configuration library. The model coefficients include spatiotemporal attention weight coefficients, domain knowledge attention coefficients, and table dimension selection coefficients.

[0011] The analysis module is used to determine the analysis report corresponding to the target analysis data and model coefficients based on the trained surface deformation analysis model. The analysis report includes deformation phenomena, deformation causes, and deformation risks.

[0012] Thirdly, an electronic device is provided, including a device processor and a memory;

[0013] The device processor is connected to the memory;

[0014] The memory is used to store executable program code;

[0015] The device processor runs a program corresponding to the executable program code stored in the memory to perform the steps of the method provided as in the first aspect or any possible implementation thereof.

[0016] Fourthly, a computer-readable storage medium is provided having a computer program stored thereon, the computer-readable storage medium storing instructions that, when executed on a computer or device processor, cause the computer or device processor to perform the method provided as in the first aspect or any possible implementation thereof.

[0017] The beneficial effects of the technical solutions provided in some embodiments of this specification include at least the following:

[0018] In one or more embodiments of this specification, an initial surface deformation analysis model is constructed, and the initial surface deformation analysis model is trained in segments based on a multi-source surface deformation point cloud to obtain a trained surface deformation analysis model. Next, the target analysis data and target analysis scenario corresponding to the received analysis command are determined using the command analysis method, and the model coefficients corresponding to the target analysis scenario are determined based on the scenario configuration library. Finally, an analysis report corresponding to the target analysis data and model coefficients is determined based on the trained surface deformation analysis model. Through the above customized training and targeted adjustments based on analysis commands, the constructed surface deformation analysis model increases its domain adaptability, meets the multi-scenario analysis needs of surface deformation point cloud data, and improves user satisfaction. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A flowchart illustrating a multi-scene analysis method for surface deformation point cloud data provided in the embodiments of this specification;

[0021] Figure 2 A schematic diagram of the structure of a multi-scene analysis device for surface deformation point cloud data provided in the embodiments of this specification;

[0022] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this specification. Detailed Implementation

[0023] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0024] The terms "first," "second," "third," etc., in the description, claims, and accompanying drawings are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such processes, methods, products, or apparatus.

[0025] The following description provides examples and does not limit the scope, applicability, or examples set forth in the claims. Changes may be made to the function and arrangement of the described elements without departing from the scope of this specification. Various processes or components may be appropriately omitted, substituted, or added to the examples. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Furthermore, features described with respect to some examples may be combined into other examples.

[0026] Please see Figure 1 , Figure 1 This document presents an overall flowchart of a multi-scene analysis method for surface deformation point cloud data provided in an embodiment of this specification.

[0027] like Figure 1 As shown, the multi-scene analysis method for this surface deformation point cloud data can include at least the following steps:

[0028] Step 101: Construct an initial surface deformation analysis model, and perform segmented training on the initial surface deformation analysis model based on the multi-source surface deformation point cloud to obtain a trained surface deformation analysis model.

[0029] The multi-source surface deformation point cloud includes a surface deformation point cloud dataset and a labeled dataset.

[0030] In the embodiments of this specification, in order to solve the problems of high dependence on manual work, difficulty in multi-source fusion, and poor scene adaptability in traditional surface deformation point cloud data analysis, an initial surface deformation analysis model can be constructed based on the Transformer architecture as the basic framework. Then, key extensions are made to the model input layer, integrating a spatiotemporal coordinate encoding module and a geographic entity embedding module. The spatiotemporal coordinate encoding module transforms the latitude, longitude, and timestamp of deformation points into high-dimensional feature vectors with spatial proximity and temporal sequence. The geographic entity embedding module establishes exclusive vector representations for professional terms such as "fault", "mining subsidence area", and "aquifer". In the model attention layer, a domain knowledge enhancement mechanism is introduced. An attention mask is generated through a preloaded geological knowledge graph to guide the model to pay more attention to semantic information strongly related to deformation during inference. Next, by statistically analyzing the surface deformation point cloud dataset and the labeled dataset, a multi-source surface deformation point cloud is constructed. The surface deformation point cloud dataset can be the original InSAR time-series observation data, with each point containing spatial coordinates, time, and deformation. The labeled dataset can be a structured set formed by geological and surveying experts to attribute and assess the deformation phenomena in specific areas. Furthermore, based on the multi-source surface deformation point cloud, the initial surface deformation analysis model is trained in stages using multimodal, domain-supervised, and reinforcement learning methods to obtain a trained surface deformation analysis model, enabling it to possess spatiotemporal correlation understanding, multi-source information fusion, and professional causal reasoning capabilities.

[0031] In one possible implementation, the step of segmenting the initial surface deformation analysis model based on a multi-source surface deformation point aggregation to obtain a trained surface deformation analysis model includes:

[0032] The fusion vector set corresponding to the surface deformation point cloud dataset is determined based on spatiotemporal coordinate encoding.

[0033] The initial surface deformation analysis model is pre-trained using the fused vector set to obtain a pre-trained surface deformation analysis model.

[0034] Based on the labeled dataset, the pre-trained surface deformation analysis model is trained using domain decision-making to obtain a well-trained surface deformation analysis model.

[0035] In the embodiments of this specification, when training the initial surface deformation analysis model in segments based on multi-source surface deformation point clouds, in order to transform the unstructured surface deformation point cloud dataset into a standardized vector sequence that the model can process, the fusion vector set corresponding to the surface deformation point cloud dataset can first be determined through spatiotemporal coordinate encoding. Specifically, the latitude and longitude in the surface deformation point cloud dataset can be normalized and periodically encoded together with the timestamp. Next, in order to enable the initial surface deformation analysis model to learn the basic features such as common spatiotemporal patterns and distribution rules of surface deformation, and to initially establish the connection between geospatial and semantic aspects, the initial surface deformation analysis model can be pre-trained using the obtained fusion vector set to obtain a pre-trained surface deformation analysis model. For example, the numerical values ​​or spatiotemporal labels of a portion of the deformation points can be randomly masked, allowing the model to make predictions based on the context. Furthermore, in order to enable the pre-trained surface deformation analysis model to acquire the ability to infer causes and risks from phenomena, it can be further trained on the pre-trained surface deformation analysis model for domain decision-making based on the labeled dataset to obtain a trained surface deformation analysis model. In the labeled dataset, each labeled data is an "input-output" pair. During training, loss functions such as cross-entropy can be optimized to enable the model to learn to infer correct professional conclusions from complex multimodal inputs.

[0036] In one possible implementation, determining the fusion vector set corresponding to the surface deformation point cloud dataset based on spatiotemporal coordinate encoding includes:

[0037] Determine the surface deformation information corresponding to each surface deformation point in the surface deformation point cloud dataset, wherein the surface deformation information includes latitude and longitude information and timestamp information;

[0038] Based on spatiotemporal coordinate encoding, the deformation feature vectors corresponding to each of the aforementioned surface deformation information are determined, and the aforementioned deformation feature vectors are integrated to obtain a fused vector set.

[0039] In the embodiments of this specification, when determining the fusion vector set corresponding to the surface deformation point cloud dataset based on spatiotemporal coordinate encoding, the surface deformation information corresponding to each surface deformation point in the surface deformation point cloud dataset can be determined first. The surface deformation information may include latitude and longitude information and timestamp information. Specifically, for a certain deformation point in the massive surface deformation point cloud data... Its corresponding latitude and longitude information and timestamp information Next, based on spatiotemporal coordinate encoding, the aforementioned surface deformation information is transformed into a deformation feature vector. Specifically, latitude and longitude values ​​can be mapped to... The interval is converted into the number of days from a certain reference time point, and finally a deformation feature vector is generated based on the spatiotemporal coordinate encoding. Finally, all deformation feature vectors The vectors are integrated by arranging them in chronological order to obtain a fused vector set.

[0040] Step 103: Determine the target analysis data and target analysis scenario corresponding to the received analysis command according to the command analysis method, and determine the model coefficients corresponding to the target analysis scenario based on the scenario configuration library.

[0041] The model coefficients include spatiotemporal attention weight coefficients, domain knowledge attention coefficients, and table dimension selection coefficients.

[0042] In the embodiments of this specification, after obtaining the trained surface deformation analysis model, in order to apply the trained surface deformation analysis model to different analysis scenarios such as "urban subsidence monitoring," "mining area safety assessment," and "landslide risk early warning," a parameterized scenario switching mechanism needs to be constructed to improve the model's practicality and applicability. Specifically, the received analysis commands can first be parsed using the semantic parsing model in the command analysis method to obtain the corresponding target analysis data and target analysis scenario. Then, the target analysis scenario is queried from the scenario configuration library to obtain its corresponding model coefficients. Among them, the model coefficients include the spatiotemporal attention weight coefficient, the domain knowledge attention coefficient, and the table dimension selection coefficient. The spatiotemporal attention weight coefficient is used to control the model's attention to the spatiotemporal evolution characteristics of the data. In the landslide monitoring scenario, this value may need to be increased to capture the spatiotemporal sequence of deformation precursors. The domain knowledge attention coefficient is used to control the degree to which the model relies on professional knowledge maps such as geology and geotechnical engineering during inference. In mining area subsidence analysis, this value is usually higher. The table dimension selection coefficient is used to determine the complexity and information dimension of the visualization tables in the output report.

[0043] In one possible implementation, determining the target analysis data and target analysis scenario corresponding to the received analysis command based on the command analysis method includes:

[0044] Receive analysis instructions and determine the analysis number and analysis area corresponding to the analysis instructions based on the Transformer parsing model;

[0045] The target analysis data corresponding to the analysis number and the target analysis scene corresponding to the analysis area are determined based on the real-time InSAR point cloud database.

[0046] In the embodiments described in this specification, the system receives analysis commands submitted by users in natural language. For example, one analysis command might be "Please generate a settlement analysis report for Area A from 2022 to the present, focusing on the area surrounding the airport." This command is input into the Transformer lightweight semantic parsing model, which identifies key entities in the command, such as "A New Area" (analysis area) and "2022 to the present" (time range). The system then understands the user's fundamental intent as "to conduct settlement analysis" and assigns a unique analysis number to this request for tracking task status and caching results. Simultaneously, the model can preliminarily determine the task's attributes based on identified features such as region, time, and deformation type. Next, the system queries the analysis number and analysis area from a real-time InSAR point cloud database. This InSAR point cloud database stores InSAR point cloud data from different regions and times. Efficient spatial and temporal indexing is then performed based on the region and time range to retrieve perfectly matching target analysis data and target analysis scenarios.

[0047] In one possible implementation, determining the model coefficients corresponding to the target analysis scenario based on the scenario configuration library includes:

[0048] The recommendation coefficient corresponding to the target analysis scenario is determined based on the scenario configuration library;

[0049] The recommendation coefficients are optimized based on feedback from the user interface to obtain the model coefficients.

[0050] In the embodiments of this specification, to make the analysis model more accurate and flexible in dealing with complex and special cases, the recommendation coefficients corresponding to the target analysis scenario can be determined first based on the scenario configuration library. The scenario configuration library is an empirical knowledge base that stores a set of recommendation coefficients obtained after extensive testing and optimization for each predefined analysis scenario. For example, for the scenario of "mining area slope stability monitoring," the library can preset the following: spatiotemporal attention weight coefficients. Domain knowledge attention coefficient Table dimension selection factor Next, the recommendation coefficient is displayed to the user through a user interface. The user can then provide feedback and optimize the recommendation coefficient based on the specific focus of the analysis, resulting in the model coefficient. Specifically, if the analysis focuses particularly on the impact of historical deformation on the present, the user can... To increase the resolution and generate a more concise report, users can... Turn it down.

[0051] Step 105: Determine the analysis report corresponding to the target analysis data and model coefficients based on the trained surface deformation analysis model.

[0052] The analysis report includes the deformation phenomenon, the causes of deformation, and the deformation risk.

[0053] In the embodiments of this specification, after obtaining the trained surface deformation analysis model, target analysis data, and model coefficients, the internal attention allocation can be adjusted according to the model coefficients to automatically analyze and reason about the input target analysis data, obtain the corresponding deformation phenomena, deformation causes, and deformation risks, and synthesize the deformation phenomena, deformation causes, and deformation risks through a preset structured report template to output the corresponding analysis report.

[0054] In one possible implementation, the step of determining the analysis report corresponding to the target analysis data and model coefficients based on the trained surface deformation analysis model includes:

[0055] Based on the model coefficients, the trained surface deformation analysis model is configured with parameters to obtain the target scene analysis model;

[0056] Based on the target scenario analysis model, determine the deformation phenomenon, deformation cause, and deformation risk corresponding to the target analysis data;

[0057] An analysis report is obtained by integrating the aforementioned deformation phenomena, causes of deformation, and deformation risks.

[0058] In the embodiments of this specification, since the trained surface deformation analysis model contains adjustable modules or parameters, the parameters of the trained surface deformation analysis model can be configured first according to the model coefficients to obtain a target scene analysis model specifically for the current analysis scenario. Next, the target analysis data is input into the configured target scene analysis model to obtain its corresponding deformation phenomena, deformation causes, and deformation risks. Furthermore, the deformation phenomena, deformation causes, and deformation risks are integrated according to a preset structured template to obtain an analysis report.

[0059] In one possible implementation, the method further includes:

[0060] In response to the deformation risk characterization as high risk in the analysis report, a risk area map corresponding to the target analysis data is generated, and the risk area map is visualized.

[0061] In the embodiments of this specification, when the deformation risk in the analysis report is characterized as high risk, in order to intuitively display the high-risk analysis results, it is necessary to extract the deformation point cloud belonging to the high-risk area in the target analysis data, and then generate a continuous risk area map through spatial interpolation. The risk area map is then visualized and interactively displayed through a graphical user interface. Users can perform operations such as zooming, panning, and clicking to query specific risk information of a point. This map is usually displayed in conjunction with the text and table parts of the analysis report on the same interface, providing a supplementary spatial perspective.

[0062] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0063] Please refer to the following. Figure 2 , Figure 2 This diagram illustrates the structure of a multi-scene analysis device for surface deformation point cloud data provided in an embodiment of this specification. It should be noted that... Figure 2 The multi-scene analysis device for surface deformation point cloud data shown is used to perform the functions described in this application. Figure 1 The methods shown in the embodiments are illustrated for ease of explanation, showing only the parts relevant to the embodiments of this application. For specific technical details not disclosed, please refer to this application. Figure 1 The example shown.

[0064] like Figure 2 As shown, the multi-scene analysis device for surface deformation point cloud data may include at least:

[0065] The construction module 201 is used to construct an initial surface deformation analysis model and to perform segmented training on the initial surface deformation analysis model based on a multi-source surface deformation point cloud to obtain a trained surface deformation analysis model. The multi-source surface deformation point cloud includes a surface deformation point cloud dataset and a labeled dataset.

[0066] The determination module 202 is used to determine the target analysis data and target analysis scenario corresponding to the received analysis command according to the command analysis method, and to determine the model coefficients corresponding to the target analysis scenario based on the scenario configuration library. The model coefficients include spatiotemporal attention weight coefficients, domain knowledge attention coefficients, and table dimension selection coefficients.

[0067] Analysis module 203 is used to determine the analysis report corresponding to the target analysis data and model coefficients based on the trained surface deformation analysis model. The analysis report includes deformation phenomena, deformation causes, and deformation risks.

[0068] In one possible implementation, the construction module 201 is specifically used for:

[0069] The fusion vector set corresponding to the surface deformation point cloud dataset is determined based on spatiotemporal coordinate encoding.

[0070] The initial surface deformation analysis model is pre-trained using the fused vector set to obtain a pre-trained surface deformation analysis model.

[0071] Based on the labeled dataset, the pre-trained surface deformation analysis model is trained using domain decision-making to obtain a well-trained surface deformation analysis model.

[0072] In one possible implementation, the construction module 201 is further configured to:

[0073] Determine the surface deformation information corresponding to each surface deformation point in the surface deformation point cloud dataset, wherein the surface deformation information includes latitude and longitude information and timestamp information;

[0074] Based on spatiotemporal coordinate encoding, the deformation feature vectors corresponding to each of the aforementioned surface deformation information are determined, and the aforementioned deformation feature vectors are integrated to obtain a fused vector set.

[0075] In one possible implementation, the determining module 202 is specifically used for:

[0076] Receive analysis instructions and determine the analysis number and analysis area corresponding to the analysis instructions based on the Transformer parsing model;

[0077] The target analysis data corresponding to the analysis number and the target analysis scene corresponding to the analysis area are determined based on the real-time InSAR point cloud database.

[0078] In one possible implementation, the determining module 202 is further configured to:

[0079] The recommendation coefficient corresponding to the target analysis scenario is determined based on the scenario configuration library;

[0080] The recommendation coefficients are optimized based on feedback from the user interface to obtain the model coefficients.

[0081] In one possible implementation, the analysis module 203 is specifically used for:

[0082] Based on the model coefficients, the trained surface deformation analysis model is configured with parameters to obtain the target scene analysis model;

[0083] Based on the target scenario analysis model, determine the deformation phenomenon, deformation cause, and deformation risk corresponding to the target analysis data;

[0084] An analysis report is obtained by integrating the aforementioned deformation phenomena, causes of deformation, and deformation risks.

[0085] In one possible implementation, the analysis module 203 is further configured to:

[0086] In response to the deformation risk characterization as high risk in the analysis report, a risk area map corresponding to the target analysis data is generated, and the risk area map is visualized.

[0087] Those skilled in the art will clearly understand that the technical solutions of the embodiments of this application can be implemented by means of software and / or hardware. In this specification, "unit" and "module" refer to software and / or hardware that can independently complete or cooperate with other components to complete a specific function, wherein the hardware may be, for example, a field-programmable gate array (FPGA), an integrated circuit (IC), etc.

[0088] Each processing unit and / or module in the embodiments of this application can be implemented by an analog circuit that implements the functions described in the embodiments of this application, or by software that executes the functions described in the embodiments of this application.

[0089] Please refer to the following. Figure 3 , Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this specification is shown.

[0090] like Figure 3 As shown, the electronic device 300 may include at least one device processor 301, at least one network interface 304, a user interface 303, a memory 305, and at least one communication bus 302.

[0091] The communication bus 302 can be used to realize the connection and communication of the above components.

[0092] The user interface 303 may include buttons, and the optional user interface may also include a standard wired interface or a wireless interface.

[0093] The network interface 304 may include, but is not limited to, Bluetooth modules, NFC modules, Wi-Fi modules, etc.

[0094] The device processor 301 may include one or more processing cores. The device processor 301 connects to various parts within the electronic device 300 using various interfaces and lines. It executes various functions and processes data of the electronic device 300 by running or executing instructions, programs, code sets, or instruction sets stored in the memory 305, and by calling data stored in the memory 305. Optionally, the device processor 301 may be implemented using at least one hardware form of DSP, FPGA, or PLA. The device processor 301 may integrate one or more of the following: CPU, GPU, and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required for display; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the device processor 301 and may be implemented as a separate chip.

[0095] The memory 305 may include RAM or ROM. Optionally, the memory 305 may include a non-transitory computer-readable medium. The memory 305 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 305 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 305 may also be at least one storage device located remotely from the aforementioned device processor 301. Figure 3 As shown, the memory 305, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and program instructions.

[0096] Specifically, the device processor 301 can be used to call a multi-scene analysis application of the surface deformation point cloud data stored in the memory 305, and specifically perform the following operations:

[0097] An initial surface deformation analysis model is constructed, and the initial surface deformation analysis model is trained in segments based on a multi-source surface deformation point cloud to obtain a trained surface deformation analysis model. The multi-source surface deformation point cloud includes a surface deformation point cloud dataset and a labeled dataset.

[0098] The target analysis data and target analysis scenario corresponding to the received analysis command are determined according to the command analysis method, and the model coefficients corresponding to the target analysis scenario are determined based on the scenario configuration library. The model coefficients include spatiotemporal attention weight coefficients, domain knowledge attention coefficients, and table dimension selection coefficients.

[0099] Based on the trained surface deformation analysis model, an analysis report corresponding to the target analysis data and model coefficients is determined. The analysis report includes deformation phenomena, causes of deformation, and deformation risks.

[0100] As an optional embodiment of this specification, the step of segmenting the initial surface deformation analysis model based on a multi-source surface deformation point cloud to obtain a trained surface deformation analysis model includes:

[0101] The fusion vector set corresponding to the surface deformation point cloud dataset is determined based on spatiotemporal coordinate encoding.

[0102] The initial surface deformation analysis model is pre-trained using the fused vector set to obtain a pre-trained surface deformation analysis model.

[0103] Based on the labeled dataset, the pre-trained surface deformation analysis model is trained using domain decision-making to obtain a well-trained surface deformation analysis model.

[0104] As an optional embodiment of this specification, the step of determining the fusion vector set corresponding to the surface deformation point cloud dataset based on spatiotemporal coordinate encoding includes:

[0105] Determine the surface deformation information corresponding to each surface deformation point in the surface deformation point cloud dataset, wherein the surface deformation information includes latitude and longitude information and timestamp information;

[0106] Based on spatiotemporal coordinate encoding, the deformation feature vectors corresponding to each of the aforementioned surface deformation information are determined, and the aforementioned deformation feature vectors are integrated to obtain a fused vector set.

[0107] As an optional embodiment of this specification, the step of determining the target analysis data and target analysis scenario corresponding to the received analysis command based on the command analysis method includes:

[0108] Receive analysis instructions and determine the analysis number and analysis area corresponding to the analysis instructions based on the Transformer parsing model;

[0109] The target analysis data corresponding to the analysis number and the target analysis scene corresponding to the analysis area are determined based on the real-time InSAR point cloud database.

[0110] As an optional embodiment of this specification, determining the model coefficients corresponding to the target analysis scenario based on the scenario configuration library includes:

[0111] The recommendation coefficient corresponding to the target analysis scenario is determined based on the scenario configuration library;

[0112] The recommendation coefficients are optimized based on feedback from the user interface to obtain the model coefficients.

[0113] As an optional embodiment of this specification, the step of determining the analysis report corresponding to the target analysis data and model coefficients based on the trained surface deformation analysis model includes:

[0114] Based on the model coefficients, the trained surface deformation analysis model is configured with parameters to obtain the target scene analysis model;

[0115] Based on the target scenario analysis model, determine the deformation phenomenon, deformation cause, and deformation risk corresponding to the target analysis data;

[0116] An analysis report is obtained by integrating the aforementioned deformation phenomena, causes of deformation, and deformation risks.

[0117] As an optional embodiment of this specification, the method further includes:

[0118] In response to the deformation risk characterization as high risk in the analysis report, a risk area map corresponding to the target analysis data is generated, and the risk area map is visualized.

[0119] This specification also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method. The computer-readable storage medium may include, but is not limited to, any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, microdrives, as well as magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic cards or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0120] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0121] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0122] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some service interface; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0123] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0124] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0125] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0126] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0127] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

Claims

1. A multi-scene analysis method for surface deformation point cloud data, characterized in that, The method includes: An initial surface deformation analysis model is constructed, and the initial surface deformation analysis model is trained in segments based on a multi-source surface deformation point cloud to obtain a trained surface deformation analysis model. The multi-source surface deformation point cloud includes a surface deformation point cloud dataset and a labeled dataset. The target analysis data and target analysis scenario corresponding to the received analysis command are determined according to the command analysis method, and the model coefficients corresponding to the target analysis scenario are determined based on the scenario configuration library. The model coefficients include spatiotemporal attention weight coefficients, domain knowledge attention coefficients, and table dimension selection coefficients. Based on the trained surface deformation analysis model, an analysis report corresponding to the target analysis data and model coefficients is determined. The analysis report includes deformation phenomena, causes of deformation, and deformation risks.

2. The method according to claim 1, characterized in that, The initial surface deformation analysis model is trained in segments based on a multi-source surface deformation point cloud to obtain a trained surface deformation analysis model, including: The fusion vector set corresponding to the surface deformation point cloud dataset is determined based on spatiotemporal coordinate encoding. The initial surface deformation analysis model is pre-trained using the fused vector set to obtain a pre-trained surface deformation analysis model. Based on the labeled dataset, the pre-trained surface deformation analysis model is trained using domain decision-making to obtain a well-trained surface deformation analysis model.

3. The method according to claim 2, characterized in that, The determination of the fusion vector set corresponding to the surface deformation point cloud dataset based on spatiotemporal coordinate encoding includes: Determine the surface deformation information corresponding to each surface deformation point in the surface deformation point cloud dataset, wherein the surface deformation information includes latitude and longitude information and timestamp information; Based on spatiotemporal coordinate encoding, the deformation feature vectors corresponding to each of the aforementioned surface deformation information are determined, and the aforementioned deformation feature vectors are integrated to obtain a fused vector set.

4. The method according to claim 1, characterized in that, The step of determining the target analysis data and target analysis scenario corresponding to the received analysis command based on the command analysis method includes: Receive analysis instructions and determine the analysis number and analysis area corresponding to the analysis instructions based on the Transformer parsing model; The target analysis data corresponding to the analysis number and the target analysis scene corresponding to the analysis area are determined based on the real-time InSAR point cloud database.

5. The method according to claim 1, characterized in that, The step of determining the model coefficients corresponding to the target analysis scenario based on the scenario configuration library includes: The recommendation coefficient corresponding to the target analysis scenario is determined based on the scenario configuration library; The recommendation coefficients are optimized based on feedback from the user interface to obtain the model coefficients.

6. The method according to claim 1, characterized in that, The step of determining the analysis report corresponding to the target analysis data and model coefficients based on the trained surface deformation analysis model includes: Based on the model coefficients, the trained surface deformation analysis model is configured with parameters to obtain the target scene analysis model; Based on the target scenario analysis model, determine the deformation phenomenon, deformation cause, and deformation risk corresponding to the target analysis data; An analysis report is obtained by integrating the aforementioned deformation phenomena, causes of deformation, and deformation risks.

7. The method according to claim 1, characterized in that, The method further includes: In response to the deformation risk characterization as high risk in the analysis report, a risk area map corresponding to the target analysis data is generated, and the risk area map is visualized.

8. A multi-scene analysis device for surface deformation point cloud data, characterized in that, The device includes: The construction module is used to construct an initial surface deformation analysis model and to perform segmented training on the initial surface deformation analysis model based on a multi-source surface deformation point cloud to obtain a trained surface deformation analysis model. The multi-source surface deformation point cloud includes a surface deformation point cloud dataset and a labeled dataset. The determination module is used to determine the target analysis data and target analysis scenario corresponding to the received analysis command according to the command analysis method, and to determine the model coefficients corresponding to the target analysis scenario based on the scenario configuration library. The model coefficients include spatiotemporal attention weight coefficients, domain knowledge attention coefficients, and table dimension selection coefficients. The analysis module is used to determine the analysis report corresponding to the target analysis data and model coefficients based on the trained surface deformation analysis model. The analysis report includes deformation phenomena, deformation causes, and deformation risks.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer-readable storage medium stores instructions that, when executed on a computer or processor, cause the computer or processor to perform the steps of the method as described in any one of claims 1-7.

Images