Electromagnetic wave data completion method and device, electronic device, and readable storage medium

Abstract

The present disclosure provides an electromagnetic wave data completion method and device, an electronic device, and a readable storage medium, which belongs to the field of data processing technology. The method includes: determining target reference data from a target reference data set based on the area type of the area to be completed. The target reference data set includes electromagnetic wave data of an existing area. The target data is completed based on the target reference data and a first mask to generate first completed data. The target data is the data to be completed, and the target data is divided into an existing area and an area to be completed. A first loss is calculated based on the first completed data and the target data, and the first completed data is corrected based on the first loss to obtain target completed data. The target completed data includes electromagnetic wave data of an existing area and electromagnetic wave data of an area to be completed. The electromagnetic wave data completion method and device, the electronic device, and the readable storage medium provided by the present disclosure can improve the accuracy of electromagnetic wave data completion.

Description

Technical Field

[0001] The present disclosure belongs to the field of data processing technology, and more specifically, relates to an electromagnetic wave data completion method and device, an electronic device, and a readable storage medium. Background Art

[0002] Geological electromagnetic wave data is data obtained by transmitting and receiving electromagnetic wave signals to underground media through various instruments (such as geological radar, etc.) in geological exploration and other activities. These data contain rich information about underground geological structure, rock properties, mineral distribution, etc. However, due to the complexity of the geological environment, instrument failure, interference factors and many other reasons, data is often missing or difficult to detect. Geological electromagnetic wave data completion is to fill and repair these missing data to obtain a relatively complete data set, so as to better serve geological research and exploration work.

[0003] However, the existing technology still has the problem of low accuracy of the completed data. Summary of the invention

[0004] The purpose of the present disclosure is to provide an electromagnetic wave data completion method and device, an electronic device, and a readable storage medium to improve the accuracy of electromagnetic wave data completion.

[0005] According to a first aspect of an embodiment of the present disclosure, there is provided an electromagnetic wave data completion method, comprising:

[0006] The target reference data is determined from the target reference data set based on the region type of the region to be completed. The target reference data set includes electromagnetic wave data of the existing region.

[0007] The target data is complemented based on the target reference data and the first mask to generate first complemented data. The target data is the data to be complemented, and the target data is divided into an existing area and an area to be complemented.

[0008] The first loss is calculated based on the first supplementary data and the target data, and the first supplementary data is corrected based on the first loss to obtain the target supplementary data. The target supplementary data includes the electromagnetic wave data of the existing area and the electromagnetic wave data of the area to be supplemented.

[0009] A second aspect of the embodiments of the present disclosure provides an electromagnetic wave data completion device, including:

[0010] The data screening module is used to determine the target reference data from the target reference data set based on the area type of the area to be completed. The target reference data set includes electromagnetic wave data of the existing area.

[0011] The first completion module is used to complete the target data based on the target reference data and the first mask to generate first completed data. The target data is the data to be completed, and the target data is divided into an existing area and a to-be-completed area.

[0012] The second complement module is used to calculate the first loss based on the first complement data and the target data, and to correct the first complement data based on the first loss to obtain the target complement data. The target complement data includes the electromagnetic wave data of the existing area and the electromagnetic wave data of the area to be complemented.

[0013] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the above-mentioned electromagnetic wave data completion method when executing the computer program.

[0014] According to a fourth aspect of the embodiments of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned electromagnetic wave data completion method are implemented.

[0015] The electromagnetic wave data completion method and device, electronic device, and readable storage medium provided by the embodiments of the present disclosure have the following beneficial effects:

[0016] On the one hand, the embodiments of the present disclosure classify the areas to be completed and screen reference data for different areas, which can ensure the data relevance and accuracy of the completion process, thereby improving the reliability of the completion results.

[0017] On the other hand, the embodiment of the present disclosure uses target reference data and a first mask to complete the target data, which not only takes into account the overall distribution characteristics of the data, but also realizes the effective retention and fusion of local information through masking technology, so that the generated first completed data is closer to the actual situation.

[0018] On the other hand, by calculating the first loss and correcting the first supplementary data, the method can continuously iterate and optimize until the target supplementary data is obtained. This process not only improves the accuracy of the supplementary data, but also ensures the consistency and coordination of the supplementary results in the entire electromagnetic wave data set.

[0019] In summary, the embodiments of the present disclosure can significantly improve the accuracy and reliability of electromagnetic wave data completion, and provide strong data support for research and applications in related fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 A schematic diagram of a flow chart of an electromagnetic wave data completion method provided in one embodiment of the present disclosure;

[0022] Figure 2 A structural block diagram of an electromagnetic wave data completion device provided in one embodiment of the present disclosure;

[0023] Figure 3 A schematic block diagram of an electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0024] In the following description, specific details such as specific system structures and technologies are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present disclosure. However, it should be clear to those skilled in the art that the present disclosure may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obstructing the description of the present disclosure with unnecessary details.

[0025] In order to make the purpose, technical solutions and advantages of the present disclosure more clear, specific embodiments will be described below in conjunction with the accompanying drawings.

[0026] Please refer to Figure 1 , Figure 1 This is a flow chart of an electromagnetic wave data completion method provided in one embodiment of the present disclosure. The method may include S101 to S103.

[0027] S101: Determine target reference data from a target reference data set based on the region type of the region to be completed. The target reference data set includes electromagnetic wave data of an existing region.

[0028] In this embodiment, the target reference data set is a database or data set storing existing regional electromagnetic wave data, including relevant data information of different regions. These data can be stored by categories such as geographic coordinates, regional characteristics, and frequency bands.

[0029] There are many types of areas to be completed, such as key areas and non-key areas according to importance, or classified by geographical features, functional uses, etc., so as to select reference data in a targeted manner. The target reference data is the data related to the area to be completed that is screened from the target reference data set. It is similar to the area to be completed in terms of regional characteristics, electromagnetic wave characteristics, etc., and the data quality meets the requirements. The electromagnetic wave data of the existing area can be obtained through monitoring by ground sensors, satellite sensors or aerial sensors.

[0030] For example, when radar detects ocean areas, some areas are missing electromagnetic wave data due to interference from waves, sea fog, etc. (areas to be completed). If the area is close to an important waterway (key area), the area type is a key area, and other existing area electromagnetic wave data similar to important waterway areas with complete data are selected from the target reference data set as target reference data. For open sea areas (non-key areas) far away from the waterway and with slightly lower accuracy requirements, the area type is a non-key area, and reference data from other corresponding open sea areas is selected to complete the missing data later and improve the integrity and accuracy of radar detection data.

[0031] S102: Completing the target data based on the target reference data and the first mask to generate first completed data. The target data is the data to be completed, and the target data is divided into an existing area and a to-be-completed area.

[0032] In this embodiment, the first mask is a binary identification matrix image, which is used to distinguish the existing area and the area to be completed in the target data. The first completed data is the data result generated after completing the target data.

[0033] In this embodiment, the Realfill technology can be used to complete the geological wave data (ie, electromagnetic wave data). Given n (n≤5) reference images (ie, target reference data) , target image (i.e. target data) and its associated binary mask (i.e. the first mask) Among them, k is the scaling factor, which is used to adjust the relationship between the reference image and the target image. In the first mask, 1 represents the area to be filled, and 0 represents The existing area in . It is expected to generate a coordinated image (i.e. the first complement data) .

[0034] When the target data is completed based on the first mask, when the mask is equal to 0, it corresponds to The pixel at this point should be kept as close as possible to Similarly, when the mask is equal to 1, it corresponds to The data at this point in the The corresponding content in . and There should be enough overlap between the contents so that a reasonable .

[0035] S103: Calculate a first loss based on the first supplementary data and the target data, and correct the first supplementary data based on the first loss to obtain target supplementary data. The target supplementary data includes electromagnetic wave data of the existing area and electromagnetic wave data of the area to be supplemented.

[0036] In this embodiment, calculating the first loss based on the first complement data and the target data includes:

[0037] A first loss of the first complement data and the target data is calculated based on the target loss function.

[0038] The objective loss function is:

[0039] )

[0040] in, , For reference data, is the target data, p is a fixed language prompt, is the element-wise product, To mask the clean image, For the first loss, is the electromagnetic wave data error measure, To complete the data points, is the measured electromagnetic wave data, For the mask, is the error value, is the regularization coefficient.

[0041] In this embodiment, for , the target loss function is calculated only for the existing area, that is, The entry of is equal to 0.

[0042] For example, text to image (Text to Image, ) diffusion repair model, injecting low-rank adaptation (Low-rankAdaptation, ) weights and use a randomly generated binary mask exist and Specifically, the open-source StableDiffusionv2 model is repaired and the LoRA layer is injected into its text encoder and U-Net for fine-tuning. p is fixed to be a sentence containing rare tokens. For each training example, multiple random rectangles can be generated and their union or the complement of the union is taken to obtain the final random mask m.

[0043] It can be concluded from the above that, on the one hand, this embodiment classifies the areas to be completed and screens reference data for different areas, which can ensure the data relevance and accuracy of the completion process, thereby improving the reliability of the completion result.

[0044] On the other hand, this embodiment uses the target reference data and the first mask to complete the target data, which not only takes into account the overall distribution characteristics of the data, but also realizes the effective retention and fusion of local information through masking technology, so that the generated first completed data is closer to the actual situation.

[0045] On the other hand, by calculating the first loss and correcting the first supplementary data, the method can continuously iterate and optimize until the target supplementary data is obtained. This process not only improves the accuracy of the supplementary data, but also ensures the consistency and coordination of the supplementary results in the entire electromagnetic wave data set.

[0046] In summary, this embodiment can significantly improve the accuracy and reliability of electromagnetic wave data completion, and provide strong data support for research and applications in related fields.

[0047] In an embodiment of the present disclosure, the region type of the region to be completed includes a key region and / or a non-key region.

[0048] Determining target reference data from a target reference data set based on the area type of the area to be completed includes:

[0049] If the area type is a key area, determine a first reference data set corresponding to the key area in the target reference data set, filter the first reference data set, and determine the target reference data. The first reference data set includes a plurality of first reference data.

[0050] If the area type is a non-key area, determine a second reference data set corresponding to the non-key area in the target reference data set, filter the second reference data set, and determine the target reference data. The second reference data set includes a plurality of second reference data.

[0051] In this embodiment, if the area type includes a key area and a non-key area, target reference data screening is performed on the key area and the non-key area respectively.

[0052] In this embodiment, the amount of data included in the first reference data is greater than the amount of data included in the second reference data. The amount of first reference data included in the first reference data set is greater than the amount of second reference data included in the second reference data set.

[0053] Exemplarily, in the scenario of geological electromagnetic wave data completion, key areas may include areas of key significance to geological exploration, engineering construction, etc., such as areas where important mineral resources, large geological structures, or major engineering construction may exist. Non-key areas are relatively less important areas in the current task, and may include areas with relatively stable geological structures and less resource distribution. The first reference data set is a large number of reference data sets for key areas, and the large amount of data included means that more abundant information can be provided for the completion of key areas. The relevant parameters of the first reference data set may include historical electromagnetic wave data of key areas, relevant geological parameters, and other data in multiple dimensions. The second reference data set is for non-key areas, and the amount of data is relatively small. The parameters may include basic electromagnetic wave characteristics and simple geological information of non-key areas. It can be screened by data matching algorithms, such as based on geographical location, similarity of geological features, etc.

[0054] For key areas, this embodiment selects from a rich first reference data set, which can utilize a large amount of detailed data to ensure the high quality of the completed data, provide a reliable basis for key geological exploration and engineering construction, and reduce the risks caused by inaccurate data, such as resource misjudgment or engineering errors. For non-key areas, a relatively simple second reference data set is used for screening, which can complete the completion without excessive consumption of resources, thereby improving the overall completion efficiency. This flexible and targeted processing method can ensure the accuracy of key area data while taking into account the completion of non-key areas, optimizing resource utilization and completion effects.

[0055] In one embodiment of the present disclosure, completing the target data based on the target reference data and the first mask includes:

[0056] If the region type is a key region, the target data is completed based on the target reference data, the first mask and the first completion accuracy.

[0057] If the area type is a non-key area, the target data is complemented based on the target reference data, the first mask and the second complement precision. The first complement precision is greater than the second complement precision.

[0058] In this embodiment, the first completion accuracy is for key areas and requires higher completion accuracy. The first completion accuracy may include parameters such as the maximum allowable error from the true value, data resolution requirements, data stability indicators, etc. For example, the maximum allowable error can be set to a smaller percentage range from the actual measurement value; the data resolution requirement can reflect the degree of change in electromagnetic wave characteristics at a specific spatial scale; the stability index can be the range limit of data fluctuations under different time or environmental conditions. The second completion accuracy means that for non-key areas, the completion accuracy requirements are relatively low. The parameters of the second completion accuracy may include a larger allowable error range from the true value than the second completion accuracy, a lower data resolution requirement than the second completion accuracy, etc.

[0059] Exemplarily, determine whether it is a key area or a non-key area based on the area type. For key areas, data is completed using target reference data, a first mask (a mark that identifies the location of the key area or the location where data is missing) and a first completion accuracy to ensure data accuracy and reliability in key areas. For non-key areas, a relatively low second completion accuracy is used to improve completion efficiency while ensuring a certain degree of rationality. Parameters can be adjusted according to different accuracy requirements by optimizing the algorithm. For key areas, a more complex algorithm model is used to complete data in combination with rich target reference data and strict accuracy requirements. For non-key areas, a simpler and faster algorithm is used to meet lower accuracy requirements and improve processing speed.

[0060] For key areas, this embodiment ensures the accuracy and reliability of data with high completion accuracy, so that relevant decisions such as geological research, resource exploration or engineering construction in the area are supported by high-quality data, avoiding major decision-making errors due to data errors. For non-key areas, the lower completion accuracy improves the completion efficiency while meeting certain rationality, which can save computing resources and time. Overall, this differentiated processing method can allocate resources more reasonably, while ensuring the data quality of key areas, and improving the efficiency and cost-effectiveness of the entire data completion work.

[0061] In one embodiment of the present disclosure, completing the target data based on the target reference data and the first mask includes:

[0062] A first mask is generated based on the target reference data.

[0063] The target data is completed based on the first mask, the target reference data and the scene knowledge of the area to be completed.

[0064] In this embodiment, the target data is completed based on the first mask, the target reference data and the scene knowledge of the area to be completed, including:

[0065] The data fluctuation value is calculated based on the target reference data.

[0066] If the data fluctuation value is greater than or equal to the data fluctuation threshold, the target data is completed based on the first mask, the target reference data and the scene knowledge of the area to be completed.

[0067] In this embodiment, scene knowledge refers to various background information about a specific geological area, including but not limited to the geological structure characteristics, rock type distribution, and the change pattern of previous electromagnetic wave data in the area. The data fluctuation value can be understood as the degree of change of the target reference data within a certain time or space range, which may include the data change amplitude, change frequency, etc. The data fluctuation threshold is a critical value used to determine whether the data fluctuation is drastic. The data fluctuation threshold can be a specific value obtained based on the statistical analysis of historical data, or different thresholds can be set according to the characteristics of different geological areas.

[0068] Exemplarily, a first mask is generated based on the target reference data. The fluctuation value of the target reference data is calculated. If the fluctuation value is greater than or equal to the threshold, it means that the data fluctuates violently. At this time, the target data is completed in combination with the first mask, the target reference data, and the scene knowledge of the area to be completed. By injecting scene knowledge, the completion model can better understand the geological wave data situation in the current environment to improve the accuracy of completion. For example, a data analysis algorithm can be used to calculate the data fluctuation value, and a reasonable data fluctuation threshold can be determined through historical data training. For the completion process, a machine learning algorithm, such as a generative adversarial network, can be used. The target reference data, the first mask, and the scene knowledge are used as inputs. After learning and optimizing the model, the completed target data is generated.

[0069] In this embodiment, due to the unpredictability of actual environmental conditions, the geological wave environment in the area where sensors are not deployed may fluctuate violently compared to the geological wave environment in the area where sensors are deployed. and There are almost no geometric constraints between them, only a few images are available as input, and the reference images have different styles, lighting conditions, and subject poses than the target. To this end, we fine-tune the pre-trained generative model by injecting scene knowledge so that the model can generate When can and To provide a conditioned understanding of the geological wave data situation in the current environment.

[0070] For example, mineral exploration is carried out in a specific geological area. For key areas, a large amount of target reference data is collected. When completing the data, the first mask is first generated based on these reference data. After calculating the data fluctuation value, it is found that it is greater than the threshold, indicating that the data fluctuates violently. At this time, combined with scene knowledge, such as the known large geological structures in the area that will affect the propagation of electromagnetic waves, and the change patterns of electromagnetic wave data in this area under similar conditions in the past, the generative adversarial network is used to complete the data, providing more accurate data support for precise mineral positioning.

[0071] This embodiment processes the characteristics of different regions to improve the pertinence and accuracy of data completion. For key areas, combined with rich target reference data, first mask and scene knowledge, it can provide more reliable data support for important activities such as mineral exploration and reduce misjudgments caused by inaccurate data. Secondly, by considering data fluctuation values ​​and thresholds, the completion strategy can be adjusted in time when the data fluctuates violently to ensure that the completion results are more in line with the actual situation. The use of advanced algorithms such as generative adversarial networks improves the efficiency and quality of data completion.

[0072] In one embodiment of the present disclosure, the first supplementary data is corrected based on the first loss, including:

[0073] The first complement data is modified based on the first loss, the second mask and the target data. The second mask is a fuzzy mask.

[0074] In this embodiment, the second mask is a fuzzy mask, which can be used to process uncertainty or fuzzy areas in the data. The first complement data is corrected by combining the first loss, the fuzzy second mask and the target data. Using the first loss as a guide, according to the correct information in the target data, and considering the fuzzy area identified by the second mask, the inaccurate part of the first complement data is adjusted, so that the complement data is closer to the real target data, and the accuracy of the complement is improved.

[0075] For example, based on the trained diffusion inpainting model, the image can be completed using the Denoising Diffusion Probabilistic Models (DDPM) sampler. (i.e. generate an image), and this generation process is based on p, and As a model condition. In practical applications, there is The existing areas in To solve this problem, the blur mask can be Process it and then use the processed synthesis and , thus obtaining the final , fully restored on the existing area and smoothly transitioned at the border of the generated area.

[0076] In addition, the diffusion inference process is random, that is, the same input conditioning image can produce any number of generated images, depending on the different input seeds of the sampling process. This randomness leads to variations in the quality of the generated results, and human intervention is required to select high-quality samples. Since the reference image can provide a basic signal for the real content of the scene and can be used to help identify high-quality outputs. The real image completion problem of geological wave data faced in this embodiment is actually a special case of the general problem statement above. Specifically, and The number of image feature correspondences between can be used as a metric to roughly quantify whether the result is faithful to the reference image. Based on this, this embodiment is based on corresponding seed selection, that is, generating a batch of output components, and obtaining { },extract and each A set of correspondences between the filled areas (i.e. The entry is equal to 1), and finally the number of matching key points is matched to the generated result { This will automatically filter out a small number of high-quality results.

[0077] By considering the first loss, this embodiment can clarify the difference direction between the completed data and the target data, effectively guide the correction process, and improve the accuracy of the completion. The second mask processes the uncertainty and fuzzy areas of the data, making the correction more targeted and adaptable to complex data situations. For complex image data such as geological wave data, this method can solve the existing regional distortion problem and perform well in recovery and transition. By sorting and screening the generated results, high-quality samples can be automatically obtained, reducing human intervention and improving work efficiency.

[0078] Corresponding to the electromagnetic wave data completion method in the above embodiment, Figure 2 This is a structural block diagram of an electromagnetic wave data completion device provided by an embodiment of the present disclosure. For ease of description, only the parts related to the embodiment of the present disclosure are shown. Figure 2 The electromagnetic wave data completion device 20 includes: a data screening module 21, a first completion module 22 and a second completion module 23.

[0079] The data screening module 21 is used to determine the target reference data from the target reference data set based on the area type of the area to be completed. The target reference data set includes electromagnetic wave data of the existing area.

[0080] The first completion module 22 is used to complete the target data based on the target reference data and the first mask to generate first completed data. The target data is the data to be completed, and the target data is divided into an existing area and a to-be-completed area.

[0081] The second supplementation module 23 is used to calculate the first loss based on the first supplementation data and the target data, and to correct the first supplementation data based on the first loss to obtain the target supplementation data. The target supplementation data includes the electromagnetic wave data of the existing area and the electromagnetic wave data of the area to be supplemented.

[0082] In one embodiment of the present disclosure, the area type of the area to be completed includes a key area and / or a non-key area. The data screening module 21 is specifically used to determine the first reference data set corresponding to the key area in the target reference data set if the area type is a key area, and screen from the first reference data set to determine the target reference data. The first reference data set includes multiple first reference data.

[0083] If the area type is a non-key area, determine a second reference data set corresponding to the non-key area in the target reference data set, filter the second reference data set, and determine the target reference data. The second reference data set includes a plurality of second reference data.

[0084] In an embodiment of the present disclosure, the first completion module 22 is specifically configured to complete the target data based on the target reference data, the first mask and the first completion accuracy if the region type is a key region.

[0085] If the area type is a non-key area, the target data is complemented based on the target reference data, the first mask and the second complement precision. The first complement precision is greater than the second complement precision.

[0086] In an embodiment of the present disclosure, the first completion module 22 is further configured to generate a first mask based on target reference data.

[0087] The target data is completed based on the first mask, the target reference data and the scene knowledge of the area to be completed.

[0088] In an embodiment of the present disclosure, the first completion module 22 is further configured to calculate a data fluctuation value based on target reference data.

[0089] If the data fluctuation value is greater than or equal to the data fluctuation threshold, the target data is completed based on the first mask, the target reference data and the scene knowledge of the area to be completed.

[0090] In one embodiment of the present disclosure, the second completion module 23 is specifically configured to calculate the first loss based on the first completion data and the target data, including:

[0091] A first loss of the first complement data and the target data is calculated based on the target loss function.

[0092] The objective loss function is:

[0093] )

[0094] in, , For reference data, is the target data, p is a fixed language prompt, is the element-wise product, To mask the clean image, For the first loss, is the electromagnetic wave data error measure, To complete the data points, is the measured electromagnetic wave data, For the mask, is the error value, is the regularization coefficient.

[0095] In an embodiment of the present disclosure, the second complement module 23 is further configured to modify the first complement data based on the first loss, the second mask and the target data. The second mask is a fuzzy mask.

[0096] See also Figure 3 , Figure 3 A schematic block diagram of an electronic device provided by an embodiment of the present disclosure. Figure 3 The electronic device 300 in the embodiment shown may include: one or more processors 301, one or more input devices 302, one or more output devices 303 and one or more memories 304. The processors 301, input devices 302, output devices 303 and memories 304 communicate with each other via a communication bus 305. The memory 304 is used to store computer programs, which include program instructions. The processor 301 is used to execute the program instructions stored in the memory 304. The processor 301 is configured to call the program instructions to execute the functions of the modules in the above-mentioned device embodiments, such as Figure 2 The functions of modules 21 to 23 are shown.

[0097] It should be understood that in the embodiment of the present disclosure, the processor 301 may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0098] The input device 302 may include a touch panel, a fingerprint collection sensor (for collecting the user's fingerprint information and fingerprint direction information), a microphone, etc., and the output device 303 may include a display (LCD, etc.), a speaker, etc.

[0099] The memory 304 may include a read-only memory and a random access memory, and provide instructions and data to the processor 301. A portion of the memory 304 may also include a non-volatile random access memory. For example, the memory 304 may also store information about the device type.

[0100] In a specific implementation, the processor 301, input device 302, and output device 303 described in the embodiments of the present disclosure can execute the implementation methods described in the first and second embodiments of the electromagnetic wave data completion method provided in the embodiments of the present disclosure, and can also execute the implementation methods of the electronic device 300 described in the embodiments of the present disclosure, which will not be repeated here.

[0101] In another embodiment of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by the processor, all or part of the processes in the above-mentioned embodiment method are implemented, and the computer program can also be completed by instructing the relevant hardware through the computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by the processor, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0102] The computer-readable storage medium may be an internal storage unit of the electronic device of any of the aforementioned embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium may also be an external storage device of the electronic device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Furthermore, the computer-readable storage medium may also include both an internal storage unit of the electronic device and an external storage device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium may also be used to temporarily store data that has been output or is to be output.

[0103] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this disclosure.

[0104] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the electronic devices and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here.

[0105] In the several embodiments provided in the present application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces or units, or it can be an electrical, mechanical or other form of connection.

[0106] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present disclosure.

[0107] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0108] The above are only specific embodiments of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present disclosure, and these modifications or replacements should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.

Claims

1. A method for completing electromagnetic wave data, characterized in that: include: Determining target reference data from a target reference data set based on an area type of the area to be completed; The target reference data set includes electromagnetic wave data of the existing area; The area type of the area to be completed includes a key area and / or a non-key area; The determining the target reference data from the target reference data set based on the area type of the area to be completed includes: If the area type is the key area, determine a first reference data set corresponding to the key area in the target reference data set, and screen the first reference data set to determine the target reference data; the first reference data set includes a plurality of first reference data; If the area type is the non-key area, determine a second reference data set corresponding to the non-key area in the target reference data set, and screen the second reference data set to determine the target reference data; the second reference data set includes a plurality of second reference data; The target data is completed based on the target reference data and the first mask to generate first completed data; the target data is data to be completed, and the target data is divided into an existing area and an area to be completed; The completing the target data based on the target reference data and the first mask includes: If the region type is the key region, completing the target data based on the target reference data, the first mask and the first completion accuracy; If the area type is the non-key area, the target data is completed based on the target reference data, the first mask and the second completion precision; the first completion precision is greater than the second completion precision; The completing the target data based on the target reference data and the first mask includes: generating the first mask based on the target reference data; Completing the target data based on the first mask, the target reference data and scene knowledge of the area to be completed; Calculating a first loss based on the first supplementary data and the target data, and correcting the first supplementary data based on the first loss to obtain target supplementary data; the target supplementary data includes the electromagnetic wave data of the existing area and the electromagnetic wave data of the area to be supplemented; The calculating a first loss based on the first supplemented data and the target data includes: Calculate a first loss of the first complementary data and the target data based on a target loss function; The objective loss function is: ） in, , For reference data, is the target data, p is a fixed language prompt, is the element-wise product, To mask the clean image, For the first loss, is the electromagnetic wave data error measure, To complete the data points, is the measured electromagnetic wave data, For the mask, is the error value, is the regularization coefficient.

2. The electromagnetic wave data completion method according to claim 1, characterized in that: The completing the target data based on the first mask, the target reference data and the scene knowledge of the area to be completed includes: Calculating a data fluctuation value based on the target reference data; If the data fluctuation value is greater than or equal to the data fluctuation threshold, the target data is completed based on the first mask, the target reference data and the scene knowledge of the area to be completed.

3. The electromagnetic wave data completion method according to claim 1, characterized in that: The correcting the first supplementary data based on the first loss includes: The first complement data is modified based on the first loss, the second mask and the target data; the second mask is a fuzzy mask.

4. An electromagnetic wave data completion device, characterized in that: include: A data screening module, used for determining target reference data from a target reference data set based on the area type of the area to be completed; The target reference data set includes electromagnetic wave data of the existing area; The area type of the area to be completed includes a key area and / or a non-key area; a data screening module, specifically configured to, if the area type is the key area, determine a first reference data set corresponding to the key area in a target reference data set, and screen the first reference data set to determine the target reference data; The first reference data set includes a plurality of first reference data; If the area type is the non-key area, determining a second reference data set corresponding to the non-key area in the target reference data set, and screening from the second reference data set to determine the target reference data; The second reference data set includes a plurality of second reference data; A first completion module, configured to complete the target data based on the target reference data and the first mask to generate first completed data; the target data is data to be completed, and the target data is divided into an existing area and an area to be completed; A first completion module, specifically configured to complete the target data based on the target reference data, a first mask and a first completion accuracy if the region type is the key region; If the area type is the non-key area, completing the target data based on the target reference data, the first mask and the second completion accuracy; The first completion precision is greater than the second completion precision; A first completion module, further configured to generate the first mask based on the target reference data; Completing the target data based on the first mask, the target reference data and scene knowledge of the area to be completed; a second complementing module, configured to calculate a first loss based on the first complementing data and the target data, and to correct the first complementing data based on the first loss to obtain target complementing data; the target complementing data includes electromagnetic wave data of the existing area and electromagnetic wave data of the area to be complemented; A second completion module, specifically configured to calculate a first loss of the first completion data and the target data based on a target loss function; The objective loss function is: ） in, , For reference data, is the target data, p is a fixed language prompt, is the element-wise product, To mask the clean image, For the first loss, is the electromagnetic wave data error measure, To complete the data points, is the measured electromagnetic wave data, For the mask, is the error value, is the regularization coefficient.

5. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

Images