A Classification and Filling Method for Aeronautical Electromagnetic Space Anomaly Coordinate Data

Abstract

The present invention discloses a method for classifying and filling abnormal coordinate data of aviation electromagnetic space, which solves problems such as unreliable processing of abnormal coordinate data of electromagnetic space. The method includes: using an isolated ellipse model fitting algorithm to find and identify abnormal values of coordinate data, and extracting abnormal coordinate data information; using an FCS random regression model for interpolation, correcting abnormal values of coordinate data, and performing data backfill; obtaining abnormal coordinate data information and an original aviation electromagnetic coordinate data set; constructing a target feature division and recognition model, and dividing features and targets for the abnormal coordinate data information and the original aviation electromagnetic coordinate data set to obtain an aviation electromagnetic coordinate data set; constructing a linear regression model, and fitting the aviation electromagnetic coordinate data after correcting the abnormal values of coordinate data to obtain corrected values of the abnormal values of coordinate data, and replacing the abnormal values of coordinate data with the corrected values of the abnormal values of coordinate data.

Description

Technical Field

[0001] The present invention relates to the technical field of airborne electromagnetic data processing, and in particular to a method for classifying and filling airborne electromagnetic space abnormal coordinate data. Background Art

[0002] Unmanned aerial vehicle (UAV) airborne electromagnetic (AE) methods are widely used in high-altitude and rugged mountainous areas, and data processing is gradually maturing. After collecting raw UAV AE data, coordinate format conversion is required in Aowei. However, this can occasionally cause anomalies in the coordinate data at certain points or locations due to factors such as satellites, the environment, and climate. Because airborne AE data involves aerial influences, it is subject to more interference from factors such as temperature, satellites, altitude, and aircraft vibration than magnetotellurics. Currently, traditional methods for processing anomaly data involve manually filling in coordinate data, which is subjective and cannot scientifically simulate and correct anomaly data. Traditional single linear regression models also struggle to accurately correct anomaly data, as single stochastic regression models cannot perform complex calculations.

[0003] Therefore, it is urgent to propose a classification and filling method for airborne electromagnetic space anomaly coordinate data that is simple in logic, accurate and reliable. Summary of the Invention

[0004] In view of the above problems, the present invention aims to provide a method for classifying and filling airborne electromagnetic space abnormal coordinate data. The technical solution adopted by the present invention is as follows:

[0005] A method for classifying and filling airborne electromagnetic space abnormal coordinate data comprises the following steps:

[0006] Step S1, obtaining the collected original airborne electromagnetic coordinate data set, using the solitary ellipse model fitting algorithm to find and identify coordinate data abnormal values; checking the data columns corresponding to the coordinate data abnormal values, and extracting abnormal coordinate data information;

[0007] Step S2, obtaining the original airborne electromagnetic coordinate data set, interpolating it using the FCS random regression model, correcting coordinate data outliers, and performing data backfilling;

[0008] Step S21, obtaining abnormal coordinate data information and original airborne electromagnetic coordinate data set;

[0009] Step S22: Construct a target feature division and recognition model, and divide the features and targets of the abnormal coordinate data information and the original airborne electromagnetic coordinate dataset to obtain an airborne electromagnetic coordinate dataset. The target feature division and recognition model includes an input layer, a backbone network, a neck network, and an output prediction layer connected in sequence. The input layer receives the abnormal coordinate data information and the original airborne electromagnetic coordinate dataset. The backbone network combines the technologies of convolutional layers and batch normalization to accelerate the training process and stabilize the model performance, and enhance the network's ability to express features and the robustness of the model. The neck network performs data processing on the enhanced airborne electromagnetic coordinate data features and uses the output prediction layer to output image data. Several parallel spatial information enhancement modules SIE are set in the neck network. The spatial information enhancement module SIE includes a coordinate convolutional layer CoordConv, a max pooling layer Maxpool, a 20th standard convolutional module CBL, a ninth convolutional layer, and a Sigmoid activation function connected in sequence, and divides the features of the spatial surrounding coordinates corresponding to the abnormal coordinate data information.

[0010] Step S23: Construct a linear regression model, fit the airborne electromagnetic coordinate data after correcting the abnormal values of the coordinate data to obtain the correction values of the abnormal values of the coordinate data, and use the correction values of the abnormal values of the coordinate data to replace the abnormal values of the coordinate data.

[0011] Furthermore, the estimated proportion of abnormal values in the original airborne electromagnetic coordinate dataset corresponding to the solitary ellipse model fitting algorithm is preset to 0.2.

[0012] Further, the backbone network includes a first standard convolution and cross-stage partial network group, a second standard convolution and cross-stage partial network group, a third standard convolution and cross-stage partial network group, a seventh standard convolution module CBL, and a spatial pyramid pooling layer connected in sequence; the first standard convolution and cross-stage partial network group includes a convolutional neural layer fcous, a first standard convolution module CBL, a first cross-stage partial network CSP1_1, a second standard convolution module CBL, a first cross-stage partial network CSP1_2, a third standard convolution module CBL, and a first cross-stage partial network CSP1_3 connected in sequence; the second standard convolution and cross-stage partial network group includes a fourth standard convolution module CBL, a second cross-stage partial network CSP1_2, a fifth standard convolution module CBL, and a second cross-stage partial network CSP1_3 connected in sequence; the third standard convolution and cross-stage partial network group includes a sixth standard convolution module CBL and a third cross-stage partial network CSP1_3 connected in sequence; the convolutional neural layer fcous is connected to the input layer; the first cross-stage partial network CSP1_3 is connected to the fourth standard convolution module CBL; the second cross-stage partial network CSP1_3 is connected to the sixth standard convolution module CBL and the neck network; the third cross-stage partial network CSP1_3 is connected to the seventh standard convolution module CBL and the neck network; the spatial pyramid pooling layer is connected to the neck network.

[0013] Further, the neck network includes a first Cross Stage Partial Network (CSP2_3), an eighth standard convolutional module (CBL), a first upsampling layer, a first concatenation channel (Concat), a second Cross Stage Partial Network (CSP2_3), a ninth standard convolutional module (CBL), a second upsampling layer, a first Spatial Information Enhancement module (SIE), a second concatenation channel (Concat), and a third Cross Stage Partial Network (CSP2_3) connected in sequence. A tenth standard convolutional module (CBL) is connected to the third Cross Stage Partial Network (CSP2_3). A second Spatial Information Enhancement module (SIE) is connected to the ninth standard convolutional module (CBL) and the tenth standard convolutional module (CBL). A third concatenation channel (Concat) is connected to the second Spatial Information Enhancement module (SIE). A fourth Cross Stage Partial Network (CSP2_3) is connected to the third concatenation channel (Concat). An eleventh standard convolutional module (CBL) is connected to the fourth Cross Stage Partial Network (CSP2_3). A third Spatial Information Enhancement module (SIE) is connected to the eleventh standard convolutional module (CBL) and the eighth standard convolutional module (CBL). A fourth concatenation channel (Concat) is connected to the third Spatial Information Enhancement module (SIE). A fifth Cross Stage Partial Network (CSP2_3) is connected to the fourth concatenation channel (Concat). The first Cross Stage Partial Network (CSP2_3) is connected to the Spatial Pyramid Pooling layer. The first concatenation channel (Concat) is connected to the third Cross Stage Partial Network (CSP1_3). The first Spatial Information Enhancement module (SIE) is connected to the second Cross Stage Partial Network (CSP1_3). The third Cross Stage Partial Network (CSP2_3), the fourth Cross Stage Partial Network (CSP2_3), and the fifth Cross Stage Partial Network (CSP2_3) are respectively connected to the output prediction layer.

[0014] Further, the structures of the first Cross Stage Partial Network (CSP1_1), the first Cross Stage Partial Network (CSP1_2), the first Cross Stage Partial Network (CSP1_3), the second Cross Stage Partial Network (CSP1_2), the second Cross Stage Partial Network (CSP1_3), and the third Cross Stage Partial Network (CSP1_3) are the same, and each includes a fourteenth standard convolutional module (CBL), a first Residual Unit (ResUnit), a fifth convolutional layer, a fifth concatenation channel (Concat), a second Batch Normalization layer, a second Leaky Rectified Linear Unit function (leakyrelu), and a fifteenth standard convolutional module (CBL) connected in sequence, and a sixth convolutional layer connected between the input of the fourteenth standard convolutional module (CBL) and the fifth concatenation channel (Concat).

[0015] Further, the structures of the first cross-stage partial network CSP2_3, the second cross-stage partial network CSP2_3, the third cross-stage partial network CSP2_3, the fourth cross-stage partial network CSP2_3, and the fifth cross-stage partial network CSP2_3 are the same, and each includes a sixteenth standard convolution module CBL, a seventeenth standard convolution module CBL, an eighteenth standard convolution module CBL, a second residual unit ResUnit, a seventh convolutional layer, a sixth connection channel Concat, a third batch normalization layer, a third rectified linear unit function leakyrelu, and a nineteenth standard convolution module CBL, as well as an eighth convolutional layer connected between the input of the sixteenth standard convolution module CBL and the sixth connection channel Concat.

[0016] Further, the structures of the first residual unit ResUnit and the second residual unit ResUnit are the same, and each includes a twelfth standard convolution module CBL, a thirteenth standard convolution module CBL, and a residual connection layer add connected in sequence.

[0017] Further, the structures of the first standard convolution module CBL, the second standard convolution module CBL, the third standard convolution module CBL, the fourth standard convolution module CBL, the fifth standard convolution module CBL, the sixth standard convolution module CBL, the seventh standard convolution module CBL, the eighth standard convolution module CBL, the ninth standard convolution module CBL, the tenth standard convolution module CBL, the eleventh standard convolution module CBL, the twelfth standard convolution module CBL, the thirteenth standard convolution module CBL, the fourteenth standard convolution module CBL, the fifteenth standard convolution module CBL, the sixteenth standard convolution module CBL, the seventeenth standard convolution module CBL, the eighteenth standard convolution module CBL, the nineteenth standard convolution module CBL, and the twentieth standard convolution module CBL are the same, and each includes a fourth convolutional layer, a first batch normalization layer, and a first rectified linear unit function leakyrelu connected in sequence.

[0018] Further, the expression of the linear regression model is:

[0019]

[0020] where represents the regression function of the linear regression model; represents the airborne electromagnetic coordinate data after correcting the outlier of the coordinate data; represents the weight; represents the bias vector.

[0021] Compared with the prior art, the present invention has the following beneficial effects:

[0022] The present invention uses the solitary ellipse model fitting algorithm to find outlier coordinate data. Since the abnormal airborne electromagnetic coordinate data is related to the adjacent airborne electromagnetic coordinate data to a certain extent, the solitary ellipse model will refer to the adjacent data, so the judgment of whether a certain point is an outlier is relatively accurate.

[0023] The present invention uses the FCS stochastic regression model to estimate the abnormal coordinate data, and randomly samples in the prediction distribution of the model to correct the abnormal coordinate values, satisfying the correlation between the abnormal points and the surrounding coordinates, and effectively avoiding the subjectivity of manually filling the abnormal coordinate data points, making the interpreted data more reliable.

[0024] The present invention uses the target feature division and recognition model to divide the features and targets of the data with missing values and the data without missing values (i.e., abnormal coordinate data information and the original airborne electromagnetic coordinate data set). Several spatial information enhancement modules SIE are set in the target feature division and recognition model, which makes the feature division of the coordinates around the abnormal data in the airborne electromagnetic spatial coordinate data more obvious, and can better reflect the spatial position points where the abnormal points are located, making the division result more reliable. After the present invention adds the spatial information enhancement module SIE, for the airborne electromagnetic spatial coordinate data, it realizes the full fusion of channel information and maintains high feature representation ability.

[0025] The spatial information enhancement module SIE of the present invention is composed of a coordinate convolution layer CoordConv, a maximum pooling layer Maxpool, a twentieth standard convolution module CBL, a ninth convolution layer, and a Sigmoid activation function. Among them, the coordinate convolution layer CoordConv modifies the standard convolution operation by adding an additional input channel encoding each pixel coordinate, which enables the convolution filter in the coordinate convolution layer CoordConv to accurately know the position in the input space, thus significantly improving the network's ability to learn spatial transformation.

[0026] (5) The present invention uses three spatial information enhancement modules SIE. When correcting data, it supplements the scene context information for the data of the abnormal points, and its output data may include the position, scale, or environmental features of the object to assist the target detection and image segmentation tasks.

[0027] In summary, the present invention has the advantages of simple logic, accuracy and reliability, etc., and has high practical value and promotion value in the field of airborne electromagnetic data processing technology. Brief Description of the Drawings

[0028] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and thus should not be regarded as limiting the protection scope. For those skilled in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0029] Figure 1 It is the logic flow chart of the present invention.

[0030] Figure 2 It is the structural schematic diagram of the input layer and the backbone network of the target feature division and recognition model in the present invention.

[0031] Figure 3 It is the structural schematic diagram of the neck network and the output prediction layer of the target feature division and recognition model in the present invention.

[0032] Figure 4 It is the structural schematic diagram of the standard convolution module CBL in the present invention.

[0033] Figure 5 It is the structural schematic diagram of the residual unit ResUnit in the present invention.

[0034] Figure 6 It is the structural schematic diagram of the cross-stage local network CSP1_X in the present invention.

[0035] Figure 7 It is the structural schematic diagram of the cross-stage local network CSP2_Y in the present invention.

[0036] Figure 8 It is the structural schematic diagram of the spatial information enhancement module SIE in the present invention. Detailed implementation manners

[0037] To make the objectives, technical solutions and advantages of the present application clearer, the following further illustrates the present invention with reference to the drawings and embodiments. The implementation manners of the present invention include but are not limited to the following embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope protected by the present application.

[0038] In this embodiment, the term "and / or" only describes the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone.

[0039] In the description of the specification and claims of this embodiment, the terms "first", "second", etc. are used to distinguish different objects, rather than to describe a specific order of the objects. For example, the first target object, the second target object, etc. are used to distinguish different target objects, rather than to describe a specific order of the target objects.

[0040] In the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design solution described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of words such as "exemplary" or "for example" is intended to present relevant concepts in a specific manner.

[0041] In the description of the embodiments of this application, unless otherwise specified, the meaning of "a plurality of" refers to two or more. For example, a plurality of processing units refers to two or more processing units; a plurality of systems refers to two or more systems.

[0042] As Figures 1 to 8 shown, this embodiment provides a method for classifying and filling abnormal coordinate data of an airborne electromagnetic space, which solves the problem of abnormal coordinate data in airborne electromagnetic space coordinate data caused by factors such as satellites, environment, and climate, and eliminates and fills the abnormal coordinate data. Specifically, it includes the following steps:

[0043] First step, obtain the collected original airborne electromagnetic coordinate data set, and use the solitary ellipse model fitting algorithm to find and identify abnormal coordinate data values; check the data columns corresponding to the abnormal coordinate data values, and extract the abnormal coordinate data information. In this embodiment, the estimated proportion of abnormal values in the original airborne electromagnetic coordinate data set corresponding to the solitary ellipse model fitting algorithm is preset to 0.2. For airborne electromagnetic space coordinate data, each coordinate is closely related to the surrounding coordinates, and the accuracy is improved by increasing the concentration value.

[0044] Second step, obtain the original airborne electromagnetic coordinate data set, perform interpolation using the FCS random regression model, correct the abnormal coordinate data values, and perform data backfilling. The specific steps are as follows:

[0045] Step S21, extract the data with missing values and the data without missing values (i.e., the abnormal coordinate data information and the original airborne electromagnetic coordinate data set).

[0046] Step S22, construct a target feature partitioning and recognition model, and partition the features and targets of the abnormal coordinate data information and the original airborne electromagnetic coordinate data set to obtain the airborne electromagnetic coordinate data set.

[0047] In this embodiment, the target feature division and recognition model includes an input layer, a backbone network, a neck network, and an output prediction layer connected in sequence. Among them, the input layer receives abnormal coordinate data information and the original airborne electromagnetic coordinate data set. In addition, the backbone network enhances the features of the airborne electromagnetic coordinate data for the abnormal coordinate data information and the original airborne electromagnetic coordinate data set. The neck network of this embodiment processes the data with enhanced features of the airborne electromagnetic coordinate data and outputs image data using the output prediction layer.

[0048] In this embodiment, the backbone network includes a first standard convolution and cross-stage local network group, a second standard convolution and cross-stage local network group, a third standard convolution and cross-stage local network group, a seventh standard convolution module CBL, and a spatial pyramid pooling layer connected in sequence. Among them, the first standard convolution and cross-stage local network group includes a convolutional neural layer fcous, a first standard convolution module CBL, a first cross-stage local network CSP1_1, a second standard convolution module CBL, a first cross-stage local network CSP1_2, a third standard convolution module CBL, and a first cross-stage local network CSP1_3 connected in sequence. In addition, the second standard convolution and cross-stage local network group includes a fourth standard convolution module CBL, a second cross-stage local network CSP1_ , a fifth standard convolution module CBL, and a second cross-stage local network CSP1_3 connected in sequence. Here, the third standard convolution and cross-stage local network group includes a sixth standard convolution module CBL and a third cross-stage local network CSP1_3 connected in sequence. In this embodiment, the second standard convolution and cross-stage local network group is the same as the second standard convolution module CBL, the first cross-stage local network CSP1_2, the third standard convolution module CBL, and the first cross-stage local network CSP1_3 in the first standard convolution and cross-stage local network group, and the third standard convolution and cross-stage local network group is the same as the fifth standard convolution module CBL and the second cross-stage local network CSP1_3 in the second standard convolution and cross-stage local network group. The advantage is to accelerate the training speed, prevent the problem of gradient disappearance, maintain the stability of the model, and enhance the feature expression ability, providing better feature information for subsequent tasks.

[0049] Here, the second standard convolution and the cross-stage local network group transmit the output abnormal coordinate data information and the original airborne electromagnetic coordinate data set to the third standard convolution and the cross-stage local network group and the first spatial information enhancement module SIE. In addition, the third standard convolution and the cross-stage local network group transmit the abnormal coordinate data information and the original airborne electromagnetic coordinate data set to the seventh standard convolution module CBL and the first connection channel Concat. In addition, the spatial pyramid delay layer in this embodiment processes the airborne electromagnetic spatial coordinate data information as follows: by fusing image features of different scales, it captures detailed information of an object at different scales, thereby improving the object recognition and positioning capabilities of the model. Detailed processing steps: 1) Construct a pyramid structure, magnify the low-scale image features and combine them with the high-scale image features; 2) Integrate information of different scales into the same channel through upsampling or downsampling operations; 3) Provide rich multi-scale description information to enhance the expressive power of the model.

[0050] In this embodiment, the structures of the first cross-stage local network CSP1_1, the first cross-stage local network CSP1_2, the first cross-stage local network CSP1_3, the second cross-stage local network CSP1_2, the second cross-stage local network CSP1_3, and the third cross-stage local network CSP1_3 are the same, and each includes a fourteenth standard convolution module CBL, a first residual unit ResUnit, a fifth convolutional layer, a fifth connection channel Concat, a second batch normalization layer, a second rectified linear unit function leakyrelu, and a fifteenth standard convolution module CBL connected in sequence, and a sixth convolutional layer connected between the input of the fourteenth standard convolution module CBL and the fifth connection channel Concat. This type of cross-stage local network CSP1_X is used to enhance the network's ability to express features and the robustness of the model. Among them, X takes values of 1, 2, and 3 respectively, corresponding to the first cross-stage local network CSP1_1, the first cross-stage local network CSP1_2, the first cross-stage local network CSP1_3, the second cross-stage local network CSP1_2, the second cross-stage local network CSP1_3, and the third cross-stage local network CSP1_3.

[0051] In this embodiment, the neck network includes a first cross-stage partial network CSP2_3, an eighth standard convolution module CBL, a first upsampling, a first connection channel Concat, a second cross-stage partial network CSP2_3, a ninth standard convolution module CBL, a second upsampling, a first spatial information enhancement module SIE, a second connection channel Concat, and a third cross-stage partial network CSP2_3 connected in sequence. A tenth standard convolution module CBL connected to the third cross-stage partial network CSP2_3, a second spatial information enhancement module SIE connected to the ninth standard convolution module CBL and the tenth standard convolution module CBL, a third connection channel Concat connected to the second spatial information enhancement module SIE, a fourth cross-stage partial network CSP2_3 connected to the third connection channel Concat, an eleventh standard convolution module CBL connected to the fourth cross-stage partial network CSP2_3, a third spatial information enhancement module SIE connected to the eleventh standard convolution module CBL and the eighth standard convolution module CBL, a fourth connection channel Concat connected to the third spatial information enhancement module SIE, and a fifth cross-stage partial network CSP2_3 connected to the fourth connection channel Concat. Herein, when the three spatial information enhancement modules SIE correct data, they supplement scene context information for the data of abnormal points, and their output data may include the position, scale, or environmental features of the object to assist in object detection and image segmentation tasks.

[0052] In this embodiment, the first spatial information enhancement module SIE, the second spatial information enhancement module SIE, and the third spatial information enhancement module SIE have the same structure, which includes a coordinate convolution layer CoordConv, a max pooling layer Maxpool, a twentieth standard convolution module CBL, a ninth convolution layer, and a Sigmoid activation function connected in sequence, and divides the features of the spatial surrounding coordinates corresponding to the abnormal coordinate data information.

[0053] In this embodiment, the first cross-stage partial network CSP2_3, the second cross-stage partial network CSP2_3, the third cross-stage partial network CSP2_3, the fourth cross-stage partial network CSP2_3, and the fifth cross-stage partial network CSP2_3 have the same structure, and all include a sixteenth standard convolution module CBL, a seventeenth standard convolution module CBL, an eighteenth standard convolution module CBL, a second residual unit ResUnit, a seventh convolution layer, a sixth connection channel Concat, a third batch normalization layer, a third leaky ReLU function leakyrelu, and a nineteenth standard convolution module CBL connected in sequence, and an eighth convolution layer connected between the input of the sixteenth standard convolution module CBL and the sixth connection channel Concat. This type of cross-stage partial network CSP2_Y is used to enhance the network's ability to express features and the robustness of the model, where Y takes the value of 3.

[0054] In this embodiment, the first residual unit ResUnit and the second residual unit ResUnit have the same structure, and both include a twelfth standard convolutional module CBL, a thirteenth standard convolutional module CBL, and a residual connection layer add that are connected in sequence. In addition, the first standard convolutional module CBL, the second standard convolutional module CBL, the third standard convolutional module CBL, the fourth standard convolutional module CBL, the fifth standard convolutional module CBL, the sixth standard convolutional module CBL, the seventh standard convolutional module CBL, the eighth standard convolutional module CBL, the ninth standard convolutional module CBL, the tenth standard convolutional module CBL, the eleventh standard convolutional module CBL, the twelfth standard convolutional module CBL, the thirteenth standard convolutional module CBL, the fourteenth standard convolutional module CBL, the fifteenth standard convolutional module CBL, the sixteenth standard convolutional module CBL, the seventeenth standard convolutional module CBL, the eighteenth standard convolutional module CBL, the nineteenth standard convolutional module CBL, and the twentieth standard convolutional module CBL have the same structure, and all include a fourth convolutional layer, a first batch normalization layer, and a first rectified linear unit function leakyrelu that are connected in sequence.

[0055] Step S23: Construct a linear regression model, fit the airborne electromagnetic coordinate data after correcting the outliers of the coordinate data to obtain the correction values of the outliers of the coordinate data, and use the correction values of the outliers of the coordinate data to replace the outliers of the coordinate data.

[0056] The above embodiments are only the preferred embodiments of the present invention, and do not limit the protection scope of the present invention. Any changes made by using the design principle of the present invention and non-creative labor on this basis shall fall within the protection scope of the present invention.

Claims

1. A method for classifying and filling abnormal coordinate data in an airborne electromagnetic space, characterized in that, It includes the following steps: Step S1: Obtain the collected original airborne electromagnetic coordinate data set, and use the solitary ellipse model fitting algorithm to find and identify coordinate data outliers; Check the data columns corresponding to the coordinate data outliers, and extract the abnormal coordinate data information; Step S2: Obtain the original airborne electromagnetic coordinate data set, perform interpolation using the FCS random regression model, correct the coordinate data outliers, and perform data backfilling; Step S21: Obtain the abnormal coordinate data information and the original airborne electromagnetic coordinate data set; Step S22: Construct a target feature division and recognition model, and divide the features and targets of the abnormal coordinate data information and the original airborne electromagnetic coordinate data set to obtain an airborne electromagnetic coordinate data set; the target feature division and recognition model includes an input layer, a backbone network, a neck network, and an output prediction layer connected in sequence; the input layer receives the abnormal coordinate data information and the original airborne electromagnetic coordinate data set; the backbone network performs coordinate data feature enhancement processing on the abnormal coordinate data information and the original airborne electromagnetic coordinate data set; the neck network performs data processing on the enhanced airborne electromagnetic coordinate data features and outputs image data using the output prediction layer; several parallel spatial information enhancement modules SIE are arranged in the neck network; the spatial information enhancement module SIE includes a coordinate convolution layer CoordConv, a maximum pooling layer Maxpool, a 20th standard convolution module CBL, a ninth convolution layer, and a Sigmoid activation function connected in sequence, and divides the features of the surrounding coordinates in the space corresponding to the abnormal coordinate data information; Step S23: Construct a linear regression model, fit the airborne electromagnetic coordinate data after correcting the coordinate data outliers to obtain a correction value for the coordinate data outliers, and use the correction value of the coordinate data outliers to replace the coordinate data outliers; the expression of the linear regression model is: , where represents the regression function of the linear regression model; represents the airborne electromagnetic coordinate data after correcting the outliers of the coordinate data; represents the weight; represents the bias vector.

2. The classification and filling method of airborne electromagnetic space anomaly coordinate data according to claim 1, characterized in that In the solitary ellipse model fitting algorithm, the estimated proportion of outliers in the original airborne electromagnetic coordinate data set is preset to 0.

2.

3. A classification and filling method for abnormal coordinate data in the aviation electromagnetic space according to claim 1 or 2, characterized in that, The backbone network includes a first standard convolution and cross-stage partial network group, a second standard convolution and cross-stage partial network group, a third standard convolution and cross-stage partial network group, a seventh standard convolution module CBL, and a spatial pyramid pooling layer connected in sequence; the first standard convolution and cross-stage partial network group includes a convolutional neural layer fcous, a first standard convolution module CBL, a first cross-stage partial network CSP1_1, a second standard convolution module CBL, a first cross-stage partial network CSP1_2, a third standard convolution module CBL, and a first cross-stage partial network CSP1_3 connected in sequence; the second standard convolution and cross-stage partial network group includes a fourth standard convolution module CBL, a second cross-stage partial network CSP1_2, a fifth standard convolution module CBL, and a second cross-stage partial network CSP1_3 connected in sequence; the third standard convolution and cross-stage partial network group includes a sixth standard convolution module CBL and a third cross-stage partial network CSP1_3 connected in sequence; the convolutional neural layer fcous is connected to the input layer; the first cross-stage partial network CSP1_3 is connected to the fourth standard convolution module CBL; the second cross-stage partial network CSP1_3 is connected to the sixth standard convolution module CBL and the neck network; the third cross-stage partial network CSP1_3 is connected to the seventh standard convolution module CBL and the neck network; the spatial pyramid pooling layer is connected to the neck network.

4. A classification and filling method for abnormal coordinate data of an airborne electromagnetic space according to claim 3, characterized in that The neck network includes a first Cross Stage Partial Network (CSP2_3), an eighth standard convolution module (CBL), a first upsampling layer, a first connection channel (Concat), a second Cross Stage Partial Network (CSP2_3), a ninth standard convolution module (CBL), a second upsampling layer, a first Spatial Information Enhancement module (SIE), a second connection channel (Concat), and a third Cross Stage Partial Network (CSP2_3) connected in sequence. A tenth standard convolution module (CBL) is connected to the third Cross Stage Partial Network (CSP2_3). A second Spatial Information Enhancement module (SIE) is connected to the ninth standard convolution module (CBL) and the tenth standard convolution module (CBL). A third connection channel (Concat) is connected to the second Spatial Information Enhancement module (SIE). A fourth Cross Stage Partial Network (CSP2_3) is connected to the third connection channel (Concat). An eleventh standard convolution module (CBL) is connected to the fourth Cross Stage Partial Network (CSP2_3). A third Spatial Information Enhancement module (SIE) is connected to the eleventh standard convolution module (CBL) and the eighth standard convolution module (CBL). A fourth connection channel (Concat) is connected to the third Spatial Information Enhancement module (SIE). A fifth Cross Stage Partial Network (CSP2_3) is connected to the fourth connection channel (Concat). The first Cross Stage Partial Network (CSP2_3) is connected to the Spatial Pyramid Pooling layer. The first connection channel (Concat) is connected to the third Cross Stage Partial Network (CSP1_3). The first Spatial Information Enhancement module (SIE) is connected to the second Cross Stage Partial Network (CSP1_3). The third Cross Stage Partial Network (CSP2_3), the fourth Cross Stage Partial Network (CSP2_3), and the fifth Cross Stage Partial Network (CSP2_3) are respectively connected to the output prediction layer.

5. A classification and filling method for abnormal coordinate data in the aviation electromagnetic space according to claim 4, characterized in that, The structures of the first Cross Stage Partial Network (CSP1_1), the first Cross Stage Partial Network (CSP1_2), the first Cross Stage Partial Network (CSP1_3), the second Cross Stage Partial Network (CSP1_2), the second Cross Stage Partial Network (CSP1_3), and the third Cross Stage Partial Network (CSP1_3) are the same, and each includes a fourteenth standard convolution module (CBL), a first Residual Unit (ResUnit), a fifth convolutional layer, a fifth connection channel (Concat), a second Batch Normalization layer, a second Leaky Rectified Linear Unit function (leakyrelu), and a fifteenth standard convolution module (CBL), as well as a sixth convolutional layer connected between the input of the fourteenth standard convolution module (CBL) and the fifth connection channel (Concat).

6. The classification and filling method for abnormal coordinate data in the aviation electromagnetic space according to claim 5, characterized in that The structures of the first cross-stage partial network CSP2_3, the second cross-stage partial network CSP2_3, the third cross-stage partial network CSP2_3, the fourth cross-stage partial network CSP2_3, and the fifth cross-stage partial network CSP2_3 are the same, and each includes a sixteenth standard convolution module CBL, a seventeenth standard convolution module CBL, an eighteenth standard convolution module CBL, a second residual unit ResUnit, a seventh convolutional layer, a sixth connection channel Concat, a third batch normalization layer, a third rectified linear unit function leakyrelu, and a nineteenth standard convolution module CBL connected in sequence, and an eighth convolutional layer connected between the input of the sixteenth standard convolution module CBL and the sixth connection channel Concat.

7. A classification and filling method for abnormal coordinate data of an airborne electromagnetic space according to claim 6, characterized in that The structures of the first residual unit ResUnit and the second residual unit ResUnit are the same, and each includes a twelfth standard convolution module CBL, a thirteenth standard convolution module CBL, and a residual connection layer add connected in sequence.

8. A method for classifying and filling abnormal coordinate data in an airborne electromagnetic space according to claim 7, characterized in that, The structures of the first standard convolution module CBL, the second standard convolution module CBL, the third standard convolution module CBL, the fourth standard convolution module CBL, the fifth standard convolution module CBL, the sixth standard convolution module CBL, the seventh standard convolution module CBL, the eighth standard convolution module CBL, the ninth standard convolution module CBL, the tenth standard convolution module CBL, the eleventh standard convolution module CBL, the twelfth standard convolution module CBL, the thirteenth standard convolution module CBL, the fourteenth standard convolution module CBL, the fifteenth standard convolution module CBL, the sixteenth standard convolution module CBL, the seventeenth standard convolution module CBL, the eighteenth standard convolution module CBL, the nineteenth standard convolution module CBL, and the twentieth standard convolution module CBL are the same, and each includes a fourth convolutional layer, a first batch normalization layer, and a first rectified linear unit function leakyrelu connected in sequence.

Images