Structural object deformation monitoring method based on Beidou spatio-temporal data fusion and AI technology

Abstract

The invention relates to the technical field of image data processing, in particular to a structure deformation monitoring method based on Beidou spatio-temporal data fusion and AI technology, comprising: acquiring a remote sensing image set of a structure under Beidou satellite time positioning and latitude and longitude positioning; performing contour edge preprocessing on each remote sensing image; determining and obtaining a key region in each structure profile diagram based on the edge change between the adjacent structure profile diagrams in the time sequence; obtaining a deformation reason of the key area corresponding to each historical time node; dividing based on the structure contour map and the key area corresponding to each remote sensing image to obtain a training group and a prediction group; training a preset neural network model based on the training group to obtain a structure deformation neural network; according to the method, the real contour map is determined after the structural body contour is extracted, the deformation key area is determined after vegetation area changes are screened out, and the deformation key area serves as input information of the structural deformation neural network, so that the accuracy of the structural deformation neural network can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and in particular to a method for monitoring structural deformation based on Beidou spatiotemporal data fusion and AI technology. Background Art

[0002] Remote sensing satellites can provide space information collection and data acquisition. Combined with the high-precision positioning technology and timing of Beidou navigation satellites, they can accurately locate the information and content displayed in remote sensing images. On this basis, combined with the support of Beidou navigation satellites for remote sensing satellite data collection, they can effectively detect and analyze the slight changes in structures over time.

[0003] However, when detecting relatively regular structures, existing methods all assume that the structure is a rectangle or a combination of rectangles. Since the acquired image may be blocked by shadows, or the morphological features of the top of the structure may be deviated and distorted due to errors in remote sensing images, it is difficult to directly obtain the regional position of the structure using a rectangle. Therefore, when using remote sensing images in the later stage, it is impossible to accurately monitor whether buildings or other structures are deformed.

[0004] Therefore, there is an urgent need for a method that can effectively use Beidou spatiotemporal data to monitor the deformation of structures. Summary of the invention

[0005] In order to solve the above technical problems, the purpose of the present invention is to provide a structural deformation monitoring method based on Beidou spatiotemporal data fusion and AI technology to solve the above problems. In order to achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0006] In a first aspect, the present application provides a method for monitoring structural deformation based on Beidou spatiotemporal data fusion and AI technology, the method comprising:

[0007] A remote sensing image set of a structure under Beidou satellite time positioning and longitude and latitude positioning is obtained, wherein the remote sensing image set includes remote sensing images corresponding to a current time node and multiple historical time nodes; contour edge preprocessing is performed on each of the remote sensing images to obtain a structure contour map corresponding to each of the remote sensing images; key areas in each of the structure contour maps are confirmed based on edge changes between the structure contour maps adjacent in time sequence, wherein the key areas are areas where contours change with time without being affected by vegetation; deformation causes of key areas corresponding to each of the historical time nodes are obtained; training groups and prediction groups are obtained based on the structure contour map and the key areas corresponding to each of the remote sensing images; a preset neural network model is trained based on the training group to obtain a structure deformation neural network; and the prediction group is input into the structure deformation neural network to obtain a deformation monitoring result of the structure.

[0008] In a possible implementation, performing contour edge preprocessing on each remote sensing image to obtain a structure contour map corresponding to each remote sensing image includes:

[0009] Based on a preset contour extraction algorithm, the remote sensing image is subjected to contour extraction and preprocessing to obtain the contour of the structure in the remote sensing image; the inflection point of the contour of the structure in the remote sensing image is identified to obtain a plurality of structural feature points; based on the structural feature points, the contour is segmented to obtain a plurality of contour edge segments; straight line analysis is performed on each of the contour edge segments, and based on the straight line analysis result, it is determined whether the contour edge segment is a structural edge line; and a structural contour map is constructed based on all the structural edge lines.

[0010] In a possible implementation, the preset contour extraction algorithm is a Canny operator edge detection algorithm.

[0011] In a possible implementation, performing straight line analysis on each of the contour edge segments, and determining whether the contour edge segment is a structure edge line based on the straight line analysis result, includes:

[0012] Based on a preset straight line recognition algorithm, straight line recognition is performed on the contour edge line segment to obtain a straight line recognition result, which includes multiple straight line segments and multiple curve segments; the slope similarity between each curve segment and its two adjacent straight line segments is calculated respectively; based on the slope similarity corresponding to each curve segment and the proportion of the contour edge line segments where all the curve segments are located, an edge line occlusion possibility evaluation value is calculated; based on the edge line occlusion possibility evaluation value and a preset first threshold, it is judged whether the contour edge line segment is a structure edge line.

[0013] In a possible implementation manner, the preset first threshold is 0.8.

[0014] In a possible implementation, the structure includes a plurality of the structures, and a structure outline map is formed based on edge lines of all the structures, including:

[0015] The slopes of the straight line segments in each of the contour edge segments are counted respectively, and the mean is calculated to obtain the mean of the slopes of the straight lines corresponding to each of the contour edge segments; the authenticity evaluation value of each of the structures is calculated based on the verticality between each pair of inflection point edge segments in each of the structures, and a pair of inflection point edge segments are two contour edge segments connected to a structural feature point; based on the authenticity evaluation value of each of the structures and a preset second threshold, it is determined whether the contour edge segment corresponding to the structure is a structural contour line; and a structural contour map is constructed based on all the structural contour lines.

[0016] In a possible implementation manner, the preset second threshold is 0.7.

[0017] In a possible implementation, the key area in each of the structure contour images is confirmed based on edge changes between the structure contour images that are adjacent in time sequence, including:

[0018] Based on the time series changes, at least one candidate edge affected by vegetation is confirmed in the structural contour map; the length of each candidate edge affected by vegetation in the time series and the tangent slope of the candidate edge affected by vegetation are counted; the evaluation index of each candidate edge affected by vegetation is calculated based on the length of each candidate edge affected by vegetation in the time series and the tangent slope of the candidate edge affected by vegetation; at least one deformed edge is obtained based on each evaluation index judgment and a third threshold judgment, and the evaluation index corresponding to the deformed edge is judged to be less than the third threshold; and a key area is constructed based on all the deformed edges.

[0019] In a possible implementation manner, the preset third threshold is 0.8.

[0020] In a possible implementation, the preset neural network model is a convolutional neural network model.

[0021] The present invention has the following beneficial effects:

[0022] In the invention, the real structure contour map is determined based on the straight line conditions of the contour and the positional relationship between the straight lines, and the vegetation area is further screened out in the area where the structure contour map has changed according to the vegetation growth over time, and the key area of ​​the structure due to deformation is determined. The deformation area is confirmed to be accurate in the remote sensing image, and the accuracy of the structure deformation neural network can be effectively improved when it is input into the structure deformation neural network. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0024] Figure 1 A flow chart of a method for monitoring structural deformation based on Beidou spatiotemporal data fusion and AI technology provided by an embodiment of the present invention;

[0025] Figure 2 A schematic diagram of a flow chart of step S2 provided in one embodiment of the present invention;

[0026] Figure 3 A schematic diagram of a result of partially recognizing a contour edge segment according to an embodiment of the present invention;

[0027] Figure 4 A schematic flow chart of step S3 provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0028] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the structural deformation monitoring method based on Beidou spatiotemporal data fusion and AI technology proposed by the present invention, its specific implementation method, structure, characteristics and effects are described in detail as follows in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0029] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0030] Example:

[0031] The following is a specific solution of a structure deformation monitoring method based on Beidou spatiotemporal data fusion and AI technology provided by the present invention, which is described in detail with reference to the accompanying drawings.

[0032] See also Figure 1 , which shows a flow chart of a method for monitoring structural deformation based on Beidou spatiotemporal data fusion and AI technology provided by an embodiment of the present invention, which includes steps S1 to S7.

[0033] S1. Obtain a remote sensing image set of a structure under Beidou satellite time positioning and longitude and latitude positioning, wherein the remote sensing image set includes remote sensing images corresponding to a current time node and multiple historical time nodes.

[0034] It should be noted that the structures mentioned in this embodiment include but are not limited to buildings, bridges, water conservancy facilities or other structures. Furthermore, in this embodiment, the components of the structure are referred to as structures. For example, the exhaust port at the top of the building, the elevator shaft and other structures. At the same time, in this embodiment, the interval between historical time nodes is preferably half a year or a year. Similarly, the time interval between the current time node and the most recent historical time node is the same as the acquisition interval of the historical time node.

[0035] S2. Performing contour edge preprocessing on each remote sensing image to obtain a structure contour map corresponding to each remote sensing image.

[0036] In this embodiment, the edge extraction of remote sensing images can be performed using the Canny operator edge detection algorithm. However, in this embodiment, it is considered that the contour edge obtained by the Canny operator edge detection may have some edges that are noise and the edge is distorted due to the influence of clouds. Therefore, see Figure 2 In this embodiment, step S2 may further specifically include steps S21 to S25, which use a single contour edge line segment as the most basic recognition unit to optimize the recognition of the distorted contour edge line segment.

[0037] First, in this embodiment, the Canny operator edge detection algorithm is still used as the preset contour extraction algorithm to perform contour extraction, and then the contour chain code tracking algorithm is used to remove noise or non-structural edge influence, see step S21 for details.

[0038] S21. Perform contour extraction and preprocessing on the remote sensing image based on a preset contour extraction algorithm to obtain the contour of the structure in the remote sensing image.

[0039] Specifically, the contour chain code tracking algorithm can be Ren Mingwu's contour chain code tracking algorithm, and the chain code of Ren Mingwu's contour chain code tracking algorithm selects an 8-neighborhood chain code structure.

[0040] S22, identifying inflection points of the contour of the structure in the remote sensing image to obtain a plurality of structural feature points.

[0041] That is, in this embodiment, the inflection point is used as the feature point of the segmentation contour. Specifically, the inflection point identification can be obtained by identifying the corner point detection algorithm, which is a prior art and will not be described in detail in this embodiment.

[0042] S23. Segment the contour based on the structural feature points to obtain a plurality of contour edge segments.

[0043] S24, performing straight line analysis on each of the contour edge line segments respectively, and determining whether the contour edge line segment is a structure edge line based on the straight line analysis result.

[0044] In this embodiment, it is considered that the edge line of the structure is generally formed by the splicing of straight lines. Therefore, in this embodiment, it is possible to directly judge whether the contour edge line segment is a straight line to determine whether the line is a structural edge line. The judgment analysis of the straight line can adopt the straight line detection algorithm in image processing, such as Hough Transform or Radon Transform. These algorithms can effectively identify the straight line features in the image, and can maintain a high accuracy rate even in the presence of noise or slight distortion. Its specific implementation process is a prior art and will not be repeated in this embodiment.

[0045] At the same time, due to the occlusion of clouds in remote sensing images, some edge segments of the contour are distorted, so that the edge segments that should be presented as straight lines may be mistakenly identified as a combination of multiple straight line segments and curve segments, such as Figure 3 The partial recognition result of the contour edge line segment shown in the figure has straight lines on its left and right sides and a curve between the two straight lines. Therefore, in order to solve the above phenomenon, the present embodiment further includes steps S241 to S244 to solve the above phenomenon in a targeted manner.

[0046] S241. Perform straight line recognition on the contour edge line segment based on a preset straight line recognition algorithm to obtain a straight line recognition result, wherein the straight line recognition result includes a plurality of straight line segments and a plurality of curve segments.

[0047] Specifically, the preset straight line recognition algorithm in this embodiment is Hough Transform or Radon Transform.

[0048] S242, respectively calculating the slope similarity between each of the curved line segments and its two adjacent straight line segments.

[0049] In this embodiment, it is considered that the curved segments may be caused by the dynamic drift of the cloud layer, which causes the straight line to be distorted into a curve. Therefore, in order to accurately identify and process these curved segments caused by occlusion, this embodiment proposes a slope similarity calculation formula:

[0050]

[0051] Among them, p u,v represents the slope similarity of the curve segment between the uth straight line segment and the vth straight line segment; k u represents the slope of the uth straight line segment; k v represents the slope of the vth straight line segment; k u,v represents the slope between the two endpoints of the curve segment between the uth straight line segment and the vth straight line segment; ε represents a constant to prevent the denominator from being zero, which can be selected by technicians in this field according to actual conditions. In this embodiment, it is preferably 0.01.

[0052] In the above calculation formula, |k u -k v | represents the difference in slope between the uth straight line segment and the vth straight line segment; It indicates the difference between the slope of the straight line segments on both sides of the curve segment and the slope of the line connecting the endpoints of the curve segment. The larger the difference is, the more the curve segment is a straight line area distorted due to occlusion.

[0053] S243: Calculate an evaluation value of the possibility of edge line occlusion based on the slope similarity corresponding to each of the curve segments and the proportion of contour edge segments where all the curve segments are located.

[0054] Specifically, in this embodiment, the edge line occlusion probability evaluation value calculation formula is as follows:

[0055]

[0056] Where n represents the total number of straight line segments within a contour edge segment; l i represents the length of the i-th straight line segment; L represents the length of the edge segment of the current contour; p j-1,j It represents the slope similarity of the curve segment between the j-1th straight line segment and the jth straight line segment; softmax() represents the activation function.

[0057] In the above calculation formula, Indicates the proportion of straight lines and uses it as a weight coefficient. The higher the proportion of straight line segments, the higher the credibility. The lower the proportion of straight line segments, the lower the credibility, so as to achieve a more accurate identification of the edge line of the structure.

[0058] S244: Determine whether the contour edge line segment is a structure edge line based on the edge line occlusion possibility evaluation value and a preset first threshold.

[0059] Specifically, in this embodiment, the first threshold is preset to be 0.8. That is, when the evaluation value of the possibility of the edge line being blocked is greater than 0.8, the contour edge line segment is actually the edge line of the structure.

[0060] In summary, this embodiment effectively solves the problem of partially distorted contour edge segments in the process of structural edge line recognition by introducing a method of analyzing the slope relationship between a curve segment and its adjacent straight line segment.

[0061] At the same time, in practice, the corner point detection algorithm can only detect points with significant changes. That is, one end or both ends of a contour edge segment may be a curve. In this case, those skilled in the art can discard this part of the curve segment.

[0062] S25, constructing a structural object contour map based on all the structural edge lines.

[0063] Furthermore, in this embodiment, it is also considered that based on the basic principles of architectural design and the stability requirements of physical structures, most of the adjacent lines of the outline of the structure in a general structure are perpendicular to each other, and only a very small number of lines may have other angles due to design factors or other factors. Therefore, in this embodiment, it is considered that most of the adjacent lines that present different angles in the remote sensing image are generally not the outline of the structure. Therefore, in this embodiment, the vertical relationship between the adjacent contour edge segments in the structure is also used as an important basis for constructing the outline map of the structure. That is, in this embodiment, step S25 can also include steps S251-step S254.

[0064] S251 , respectively counting the slopes of the straight line segments within each of the contour edge line segments, and performing mean calculation to obtain the mean of the slopes of the straight lines corresponding to each of the contour edge line segments.

[0065] S252. Obtain an authenticity evaluation value of each structure based on the perpendicularity between each pair of inflection point edge line segments in each structure, wherein a pair of inflection point edge line segments are two contour edge line segments connected to a structural feature point.

[0066] Specifically, in this embodiment, the calculation formula of the authenticity evaluation value is as follows:

[0067]

[0068] Among them, y represents the authenticity evaluation value, The angle value corresponding to the mean value of the slope of the straight line of the i-th contour edge segment in the structure; It represents the angle value corresponding to the mean value of the straight line slope of the i+1th contour edge segment in the structure; m represents the number of contour edge lines that the structure is divided by structural feature points.

[0069] In the above calculation formula, Indicates whether the angle between the two contour edge segments is 90 degrees. If it is 90 degrees, the sin value is 1. If it is approximately 90 degrees, the sin value is also close to 1.

[0070] S253: determining whether the contour edge line segment corresponding to each of the structures is a structure contour line based on the authenticity evaluation value of each of the structures and a preset second threshold.

[0071] Specifically, in this embodiment, the second threshold is preset to be 0.7. That is, when the authenticity evaluation value of the structure is greater than 0.8, the contour edge line segment corresponding to the structure is actually the contour line of the structure.

[0072] S254: construct a structure contour map based on all the structure contour lines.

[0073] That is, each structure is analyzed and judged one by one in the above manner. Finally, the contour edge segments corresponding to multiple structures form the structure contour map, thereby reducing the impact of remote sensing image distortion.

[0074] Through the above steps, when extracting the outline of a structure not affected by vegetation, this embodiment first extracts the outline of the structure, and then determines the straight line conditions of the inner contour edge lines of the structure and the positional relationship between the straight lines to determine the true structure outline.

[0075] S3. Confirming a key area in each structure outline image based on edge changes between the structure outline images that are adjacent in time sequence, wherein the key area is an area where the outline changes with time without being affected by vegetation.

[0076] In this embodiment, it is considered that as the time nodes change, the differences between different time nodes are not only the result of the deformation of the structure, but also include the impact of the accompanying growth of vegetation. Therefore, in this embodiment, a method for further identifying key areas is included. For details, please refer to Figure 4 , which shows that step S3 includes steps S31 to S35.

[0077] S31. Confirm and obtain at least one candidate vegetation impact edge in the structure contour map based on time series changes.

[0078] In this embodiment, the method for confirming the candidate edge affected by vegetation is to compare the contour edge line segments of the same position at different time nodes one by one to see whether they have changed. If there is a change, it is the candidate edge affected by vegetation.

[0079] S32: Count the length of each candidate vegetation impact edge in the time series and the maximum tangent slope of the candidate vegetation impact edge.

[0080] It should be noted that the length of each vegetation impact candidate edge in the time series mentioned in this step refers to the length of each vegetation impact candidate edge at each historical time node. At the same time, the maximum tangent slope of the vegetation impact candidate edge refers to the maximum value of the tangent slope of the curve part of the vegetation impact candidate edge.

[0081] S33. Calculate an evaluation index of each candidate edge for vegetation impact based on the length of each candidate edge for vegetation impact in the time series and the tangent slope of the candidate edge for vegetation impact.

[0082] It should be noted that the evaluation index calculation formula is as follows:

[0083]

[0084] Among them, w represents the evaluation index; s represents the total number of time nodes; lq,q+# represents the length of the candidate edge affected by the vegetation changes from the qth historical time node to the q+1th historical time node; θ q,q+# It represents the angle difference between the maximum tangent slope corresponding to the qth historical time point and the maximum tangent slope corresponding to the q+1th historical time point, which can also be called the angle difference from the qth historical time point to the q+1th historical time point; It represents the average angle of the maximum tangent slope of multiple adjacent differences of the edge. The smaller it is, the more likely it is that the vegetation affects the candidate edge changes extending on a line. ε represents a constant to prevent the denominator from being zero, and can be selected by those skilled in the art according to actual conditions. In this embodiment, it is preferably 0.01.

[0085] S34. Obtain at least one deformed edge based on each of the evaluation index judgments and the third threshold judgment, wherein the evaluation index judgment corresponding to the deformed edge is less than the third threshold.

[0086] It should be noted that, in this embodiment, the third threshold is preferably 0.8.

[0087] S35, constructing a key area based on each of the deformed edges.

[0088] It should be noted that the key area mentioned in this step is surrounded by several small boxes, each of which includes a key area. The size of each small box is determined according to the scale of the remote sensing image, and no specific limitation is made in this embodiment.

[0089] S4. Obtain the deformation cause of the key area corresponding to each historical time node.

[0090] Specifically, in this embodiment, the deformation cause is manually marked, and the deformation cause may be displacement, settlement, tilt, no change, etc.

[0091] S5. Dividing the structure contour map and the key area corresponding to each remote sensing image into a training group and a prediction group.

[0092] S6. Train the preset neural network model based on the training group to obtain a structural deformation neural network.

[0093] Specifically, in this embodiment, the preset neural network model is a convolutional neural network model. Meanwhile, other network models may also be selected by those skilled in the art, and no specific limitation is made in this embodiment.

[0094] At the same time, the training process mentioned in this step is to divide the training group into training samples and verification samples in a ratio of 7:3 or 8:2 to train and verify the preset neural network model. And in an embodiment, the prediction group includes a structural outline map and a key area corresponding to the current time node and at least one historical time node. The number of time nodes included in the prediction group is determined by the logical computing power of the carrier implemented by the technicians in this field, and no specific restrictions are made in this embodiment. Correspondingly, each sample in the training sample and the verification sample is composed of a structural outline map and a key area with the same number of time nodes in the prediction group, and the time nodes in each sample are adjacent to each other.

[0095] S7. Inputting the prediction group into the structural deformation neural network to obtain a deformation monitoring result of the structure.

[0096] In this embodiment, the outline of the structure is first extracted by using the Canny operator edge detection, and then the real structure outline is determined based on the straight line conditions of the outline and the change state of the curve segments between the straight line segments, which can effectively reduce the dynamic impact of cloud cover. Finally, by comparing the changes in the structure outline at different historical time nodes, and based on the vegetation growth over time, the vegetation area is screened out for the areas where there are changes in the structure outline, so as to determine the key areas of the structure due to deformation. Finally, it is input into the structure deformation neural network to effectively improve the prediction accuracy.

[0097] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0098] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for monitoring structural deformation based on Beidou spatiotemporal data fusion and AI technology, characterized in that: The method comprises: Acquire a remote sensing image set of the structure under Beidou satellite time positioning and longitude and latitude positioning, wherein the remote sensing image set includes remote sensing images corresponding to a current time node and multiple historical time nodes; Performing contour edge preprocessing on each of the remote sensing images to obtain a structure contour map corresponding to each of the remote sensing images; Confirming a key area in each of the structure contour images based on edge changes between the structure contour images that are adjacent in time sequence, wherein the key area is an area where the contour changes with time without being affected by vegetation; Obtaining deformation causes of key areas corresponding to each of the historical time nodes; Dividing the structure contour map and the key area corresponding to each remote sensing image into a training group and a prediction group; Based on the training group, a preset neural network model is trained to obtain a structural deformation neural network; The prediction group is input into the structural deformation neural network to obtain the deformation monitoring result of the structure.

2. The method for monitoring structural deformation based on Beidou spatiotemporal data fusion and AI technology according to claim 1 is characterized in that: Performing contour edge preprocessing on each of the remote sensing images to obtain a structure contour map corresponding to each of the remote sensing images includes: Performing contour extraction and preprocessing on the remote sensing image based on a preset contour extraction algorithm to obtain the contour of the structure in the remote sensing image; Performing inflection point recognition on the contour of the structure in the remote sensing image to obtain a plurality of structural feature points; Segmenting the contour based on the structural feature points to obtain a plurality of contour edge segments; Performing straight line analysis on each of the contour edge segments respectively, and judging whether the contour edge segment is a structure edge line based on the straight line analysis result; A structure contour map is constructed based on all the structure edge lines.

3. The method for monitoring structural deformation based on Beidou spatiotemporal data fusion and AI technology according to claim 2 is characterized in that: The preset contour extraction algorithm is the Canny operator edge detection algorithm.

4. The method for monitoring structural deformation based on Beidou spatiotemporal data fusion and AI technology according to claim 2 is characterized in that: Performing straight line analysis on each of the contour edge segments respectively, and determining whether the contour edge segment is a structure edge line based on the straight line analysis result, including: Based on a preset straight line recognition algorithm, the contour edge line segment is subjected to straight line recognition to obtain a straight line recognition result, wherein the straight line recognition result includes a plurality of straight line segments and a plurality of curved line segments; respectively calculating the slope similarity between each of the curve segments and its two adjacent straight line segments; The possibility evaluation value of edge line being blocked is calculated based on the slope similarity corresponding to each of the curve segments and the proportion of the contour edge segments where all the curve segments are located; Based on the edge line occlusion possibility evaluation value and a preset first threshold, it is determined whether the contour edge line segment is a structure edge line.

5. The method for monitoring structural deformation based on Beidou spatiotemporal data fusion and AI technology according to claim 4 is characterized in that: The preset first threshold is 0.

8.

6. The method for monitoring structural deformation based on Beidou spatiotemporal data fusion and AI technology according to claim 2 is characterized in that: The structure includes a plurality of the structures, and a structure outline is formed based on edge lines of all the structures, including: The slopes of the straight line segments in each of the contour edge line segments are counted respectively, and the mean is calculated to obtain the mean value of the straight line slopes corresponding to each of the contour edge line segments; The authenticity evaluation value of each structure is obtained by calculating the verticality between each pair of inflection point edge line segments in each structure, wherein a pair of inflection point edge line segments are two contour edge line segments connected to a structural feature point; Based on the authenticity evaluation value of each of the structures and a preset second threshold, it is determined whether the contour edge line segment corresponding to the structure is a structure contour line; A structure contour map is constructed based on all the structure contour lines.

7. The method for monitoring structural deformation based on Beidou spatiotemporal data fusion and AI technology according to claim 1 is characterized in that: The preset second threshold is 0.

7.

8. The method for monitoring structural deformation based on Beidou spatiotemporal data fusion and AI technology according to claim 1 is characterized in that: The key area in each structure outline image is confirmed based on the edge changes between the structure outline images that are adjacent in time sequence, including: Confirming at least one candidate vegetation impact edge in the structure outline map based on the temporal changes; Count the length of each candidate edge affected by vegetation in the time series and the maximum tangent slope of the candidate edge affected by vegetation; Calculating an evaluation index of each candidate edge for vegetation impact based on the length of each candidate edge for vegetation impact in the time series and the tangent slope of the candidate edge for vegetation impact; Obtaining at least one deformed edge based on each of the evaluation index judgments and the third threshold judgment, respectively, wherein the evaluation index judgment corresponding to the deformed edge is less than the third threshold; A key area is constructed based on each of the deformed edges.

9. The method for monitoring structural deformation based on Beidou spatiotemporal data fusion and AI technology according to claim 8 is characterized in that: The preset third threshold is 0.

8.

10. The method for monitoring structural deformation based on Beidou spatiotemporal data fusion and AI technology according to claim 1, characterized in that: The preset neural network model is a convolutional neural network model.

Images