Ground surface settlement monitoring method, device and equipment based on data fusion

Abstract

The invention discloses a ground surface settlement monitoring method, device and equipment based on data fusion, and relates to the field of mining area safety management, and the method comprises the steps: obtaining a settlement value observed by adopting UAV-L DAR and a settlement value observed by adopting SBAS-I < nSAR > in a settlement area; obtaining a coherence coefficient according to the SBAS-I nSAR data, and dividing a settlement area based on the coherence coefficient; the settlement value observed by SBAS-I nSAR is adopted as the final deformation value of the high-coherence area, and the settlement value measured by UAV-L i DAR is adopted as the final deformation value of the low-coherence area; determining a corresponding weight according to the value of the coherence coefficient of each point in the transition area, fusing settlement values observed by adopting UAV-L i DAR and SBAS-I nSAR to obtain a settlement value after weight fusion, and optimizing the settlement value after weight fusion by combining a regression coefficient to obtain settlement information after regression optimization of the transition area; and performing linear superposition on the final deformation values of the high-coherence area and the low-coherence area and the settlement information after regression optimization of the transition area to obtain a final settlement field of the settlement basin.

Description

Technical Field

[0001] The present invention relates to the technical field of mine safety management, and particularly to a surface subsidence monitoring method, device and equipment based on data fusion. Background Art

[0002] Surface subsidence monitoring based on data fusion in goaf areas is an important research content in mine safety management. To comprehensively and precisely obtain surface subsidence information, two technologies, SBAS-InSAR and UAV-LiDAR, provide their respective unique advantages:

[0003] The SBAS-InSAR technology has the ability to cover a large area and is suitable for continuous settlement monitoring of long time series. Especially in areas with small-gradient settlement, its monitoring accuracy is relatively high. However, in areas with large-gradient settlement, due to the settlement rate may exceed the maximum detectable deformation gradient of InSAR technology, its monitoring ability is limited. The UAV-LiDAR technology is characterized by high resolution and high precision, and can carefully capture the large-gradient settlement information in the center area of the subsidence basin. However, in areas with small-gradient settlement, its accuracy is relatively low, and at the same time, limited by the data acquisition cost and operation coverage, it is difficult to conduct large-area continuous monitoring.

[0004] Traditional accuracy evaluation methods mainly rely on selecting multiple observation points inside the subsidence basin and comparing their settlement values with high-precision reference data (such as leveling measurement, RTK data), and analyzing the overall accuracy through error statistics. However, the shape and settlement characteristics of the subsidence basin often have obvious spatial differences, such as the "funnel-shaped" distribution characteristics with large settlement in the center and small settlement at the boundary, which makes the monitoring accuracy requirements in different regions vary.

[0005] In large-scale coal mining areas, the subsidence basin often shows a trend of large settlement in the center and small settlement at the boundary. The settlement amplitude in the center area of the basin may reach several meters to dozens of meters. At this time, even if the monitoring accuracy is within the decimeter level, it is sufficient to meet the settlement monitoring requirements. In the boundary area of the basin, since the settlement is usually only centimeter-level or even millimeter-level, the monitoring accuracy needs to be improved accordingly, otherwise it may lead to misjudgment of the settlement information in the boundary area. Therefore, a single error evaluation method is difficult to comprehensively reflect the applicability of the fused data, resulting in insufficient accuracy of surface subsidence monitoring based on data fusion. Summary of the Invention

[0006] The present invention provides a surface subsidence monitoring method, device and equipment based on data fusion to solve the above problems existing in the prior art, that is, how to improve the accuracy of surface subsidence monitoring based on data fusion in the prior art. The present invention provides a surface subsidence monitoring method based on data fusion, and the method includes:

[0007] Obtain the settlement values observed by UAV-LiDAR and the settlement values observed by SBAS-InSAR within the settlement area in the same time dimension;

[0008] Based on the obtained SBAS-InSAR data, obtain the coherence coefficient within the settlement area, and divide the settlement area based on the coherence coefficient to obtain a high coherence area, a low coherence area, and a transition area;

[0009] Use the settlement value observed by SBAS-InSAR as the final deformation value of the high coherence area, and use the settlement value measured by UAV-LiDAR as the final deformation value of the low coherence area;

[0010] According to the coherence coefficient value of each point in the transition area, determine the corresponding weight, fuse the settlement values observed by UAV-LiDAR and SBAS-InSAR for each point in the transition area to obtain the weighted fused settlement value, and based on the weighted fused settlement value, combine the regression coefficient determined by the least squares method OLS, and optimize the weighted fused settlement value through the regression coefficient to obtain the settlement information after regression optimization in the transition area;

[0011] Linearly superimpose the final deformation value of the high coherence area, the final deformation value of the low coherence area, and the settlement information after regression optimization in the transition area to obtain the final settlement field of the subsidence basin.

[0012] Optionally, the division of the settlement area based on the coherence coefficient specifically includes:

[0013] According to the coherence coefficient, use the following formula to obtain the maximum detectable deformation gradient:

[0014]

[0015] where d is the maximum detectable deformation gradient, γ is the coherence coefficient, λ and μ are the radar wavelength and the pixel or spatial resolution of a single SAR image respectively;

[0016] Divide the settlement area according to the relationship between the coherence coefficient and the detectable deformation gradient.

[0017] Optionally, the process of determining the corresponding weight according to the coherence coefficient value of each point in the transition area, fusing the settlement values observed by UAV-LiDAR and SBAS-InSAR for each point in the transition area to obtain the weighted fused settlement value, and based on the weighted fused settlement value, combining the regression coefficient determined by the least squares method OLS, and optimizing the weighted fused settlement value through the regression coefficient to obtain the settlement information after regression optimization in the transition area specifically includes:

[0018] According to the coherence coefficient values of each point in the transition region, the weights of the InSAR monitoring results and the UAV-LiDAR monitoring results are obtained using the following formula:

[0019]

[0020] W UAV = 1 - W InSAR

[0021] where r ave represents the average coherence value in the transition region, r i is the coherence value of the i-th point in the transition region, W InSAR represents the weight of the InSAR monitoring results, and W UAV represents the weight of the UAV-LiDAR monitoring results;

[0022] According to the weights of the InSAR monitoring results and the UAV-LiDAR monitoring results, the weighted fusion settlement value is obtained using the following formula:

[0023] D(x i , y i ) = W InSAR D InSAR (x i , y i ) + W UAV D UAV (x i , y i )

[0024] where (x i , y i ) represents the coordinates of the i-th point, D InSAR (x i , y i ) represents the settlement value of the i-th point observed by SBAS-InSAR, D UAV (x i , y i ) represents the settlement value of the i-th point measured by UAV-LiDAR, and D'(x i , y i ) represents the weighted fusion settlement value of the i-th point;

[0025] According to the weighted fusion settlement value, combined with the regression coefficients determined by the least squares method OLS, the weighted fusion settlement value is regressively optimized and estimated through the regression coefficients, and the following formula is used to obtain the settlement information after regression optimization in the transition region:

[0026] D' fin (x i , y i ) = β0(x i, y i ) + β1(x i , y i )D'(x i , y i ) + ε i

[0027] Among them, D′ fin (x i , y i ) is the final fusion settlement value of the transition area, and β0(x i , y i ) and β1(x i , y i ) are both the regression coefficients corresponding to the i-th point, and ε i is the error term, where i = 1, 2,..., n, and n is the number of discrete points.

[0028] Optionally, the linear superposition of the final deformation value of the high-coherence area, the final deformation value of the low-coherence area, and the settlement information after regression optimization of the transition area to obtain the final subsidence basin settlement field specifically includes:

[0029] The following formula is used to obtain the ground settlement value to be estimated:

[0030]

[0031] Among them, d is the ground settlement value to be estimated, D' fin (x i , y i ) is the settlement information after regression optimization of the transition area, where i = 1, 2,..., n, and n is the number of discrete points, D InSAR (x i , y i ) represents the settlement value observed by SBAS-InSAR at the i-th point, and D UAV (x i , y i ) represents the settlement value measured by UAV-LiDAR at the i-th point.

[0032] Optionally, the acquisition of the settlement values observed by UAV-LiDAR and the settlement values observed by SBAS-InSAR in the same time dimension within the settlement area specifically includes:

[0033] Resample and perform coordinate transformation on the settlement values observed by UAV-LiDAR so that the settlement values observed by UAV-LiDAR are the same as the settlement values observed by SBAS-InSAR in terms of spatio-temporal resolution and coordinate system.

[0034] The present invention provides a surface settlement monitoring device based on data fusion, including:

[0035] An acquisition module, configured to acquire the settlement values observed by UAV-LiDAR and the settlement values observed by SBAS-InSAR within the same time dimension in the settlement area;

[0036] A division module, configured to obtain the coherence coefficient within the settlement area according to the acquired SBAS-InSAR data, divide the settlement area based on the coherence coefficient to obtain a high coherence area, a low coherence area, and a transition area;

[0037] A determination module, configured to use the settlement value observed by SBAS-InSAR as the final deformation value of the high coherence area, and use the settlement value measured by UAV-LiDAR as the final deformation value of the low coherence area; according to the coherence coefficient value of each point in the transition area, determine the corresponding weight, fuse the settlement values observed by UAV-LiDAR and SBAS-InSAR for each point in the transition area to obtain the weighted fused settlement value, and according to the weighted fused settlement value, combine the regression coefficient determined by the least squares method OLS to optimize the weighted fused settlement value through the regression coefficient to obtain the settlement information after regression optimization in the transition area;

[0038] A fusion module, configured to linearly superimpose the final deformation value of the high coherence area, the final deformation value of the low coherence area, and the settlement information after regression optimization in the transition area to obtain the final settlement field of the subsidence basin.

[0039] The present invention provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the above-mentioned surface subsidence monitoring method based on data fusion is implemented.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention provides a surface subsidence monitoring method based on data fusion. By establishing the relationship between coherence and detectable deformation gradient, high coherence regions, low coherence regions, and transitional regions are obtained. Through a weighted fusion method based on coherence weights, weighted fusion is performed within the transitional regions to obtain the subsidence field after fusion in the transitional regions, which can make full use of the stability of the SBAS-InSAR technology in small-gradient subsidence regions and the high-precision characteristics of the UAV-LiDAR technology in large-gradient subsidence regions to improve the monitoring accuracy in the transitional regions. In addition, by using a regression model to optimize the subsidence field after fusion in the transitional regions, local errors can be reduced, making the subsidence basin smoother and more reasonable, and thus obtaining more complete subsidence information. Finally, by linearly superimposing the high coherence regions, low coherence regions, and the transitional regions after regression optimization, a complete and detailed surface subsidence field of the goaf is constructed, improving the accuracy and reliability of surface subsidence monitoring based on data fusion. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings herein are incorporated into the specification and form a part of this specification, showing embodiments consistent with the present invention, and are used together with the specification to explain the principles of the present invention.

[0042] Figure 1 It is a flowchart of a surface subsidence monitoring method based on data fusion provided by an embodiment of the present invention;

[0043] Figure 2 It is a schematic diagram of a fusion strategy provided by an embodiment of the present invention;

[0044] Figure 3 It is a flowchart of a fusion basin provided by an embodiment of the present invention;

[0045] Figure 4 It is a schematic diagram of the full-basin data fusion result of UAV-LiDAR and SBAS-InSAR monitoring provided by an embodiment of the present invention;

[0046] Figure 5 It is a map of the division of the surface subsidence basin provided by an embodiment of the present invention;

[0047] Figure 6 It is a statistical chart of the errors of the fused subsidence basin provided by an embodiment of the present invention;

[0048] Figure 7 It is a schematic diagram of a computer device for a surface subsidence monitoring method based on data fusion provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0049] To make the objectives, technical solutions, and advantages of the present invention more clear, the following will clearly and completely describe the technical solutions in the present invention in combination with the accompanying drawings in the present invention. Apparently, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0050] The following uses specific embodiments to detail the technical solutions of the present invention and how the technical solutions of the present invention solve the above technical problems. The following several specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The following will describe the embodiments of the present invention in combination with the accompanying drawings.

[0051] Figure 1 is a flowchart of a ground settlement monitoring method based on data fusion provided by an embodiment of the present invention. As Figure 1 shown, a ground settlement monitoring method based on data fusion shown in this embodiment includes:

[0052] S1: Obtain the settlement values observed by UAV-LiDAR and the settlement values observed by SBAS-InSAR within the settlement area in the same time dimension.

[0053] Exemplarily, obtain the UAV-LiDAR subsidence basin at the same time interval to ensure its consistency with the SBAS-InSAR data in the time dimension. The resampling operation can be performed on the obtained UAV-LiDAR subsidence basin data by using ArcGIS to make its spatial resolution consistent with the SBAS-InSAR data, and convert the coordinate system of the UAV-LiDAR subsidence basin data to the same coordinate system as the SBAS-InSAR data. For example, the coordinate system of the UAV-LiDAR subsidence basin data can be converted to the WGS84 coordinate system to be consistent with the SBAS-InSAR subsidence basin data. Furthermore, three groups of data, namely the SBAS-InSAR settlement value, the UAV-LiDAR settlement value, and the coherence value for fusion, are obtained, laying a foundation for subsequent regional division and weight calculation.

[0054] S2: According to the obtained SBAS-InSAR data, obtain the coherence coefficient within the settlement area, and divide the settlement area based on the coherence coefficient to obtain a high-coherence area, a low-coherence area, and a transition area.

[0055] Generally, according to the prior art, the functional relationship between the detectable deformation gradient and the coherence coefficient of the SAR image is:

[0056]

[0057] Among them, d is the maximum detectable deformation gradient of the InSAR technology, γ is the coherence coefficient; λ and μ are the radar wavelength and the pixel or spatial resolution of a single SAR image respectively. According to the principle of Equation 4-1, it can be obtained that: the larger the coherence coefficient γ, the smaller the corresponding detectable deformation gradient d; on the contrary, the smaller the coherence coefficient γ, the larger the corresponding detectable deformation gradient d. This characteristic can be used to divide regions with different coherences, namely high coherence regions, low coherence regions, and transition regions.

[0058] To further quantify the division boundaries of different regions, the maximum coherence coefficient γ is defined upper and the minimum coherence coefficient γ lower , and the upper and lower bounds of the transition region are calculated according to the following formula:

[0059]

[0060] Among them, d upper represents the maximum detectable deformation gradient of the region with higher coherence, while d lower represents the maximum detectable deformation gradient of the region with lower coherence. Through this method, the spatial ranges of different regions can be reasonably delimited, thereby providing a theoretical basis for the fusion of InSAR and UAV-LiDAR data, as Figure 2 shown. The flow chart of monitoring the mining subsidence basin based on the fusion of SBAS-InSAR and UAV-LiDAR monitoring technologies is as Figure 3 shown.

[0061] S3: Use the settlement value observed by SBAS-InSAR as the final deformation value of the high coherence region, and use the settlement value measured by UAV-LiDAR as the final deformation value of the low coherence region.

[0062] S4: According to the value of the coherence coefficient of each point in the transition region, determine the corresponding weight, fuse the settlement values observed by UAV-LiDAR and SBAS-InSAR for each point in the transition region to obtain the weighted fused settlement value, and according to the weighted fused settlement value, combined with the regression coefficient determined by the least squares method OLS, optimize the weighted fused settlement value through the regression coefficient to obtain the settlement information after regression optimization in the transition region.

[0063] Exemplarily, in the transition region, the method of dynamic weight allocation of coherence is used to balance the contributions of the two data sources. Its weight calculation formula is as follows:

[0064]

[0065] W UAV = 1 - W InSAR

[0066] Among them, r ave represents the average coherence value in the transition region, and r i is the coherence value at the i-th point in the transition region. W InSAR represents the weight of the InSAR monitoring result, and W UAV represents the weight of the UAV-LiDAR monitoring result;

[0067] According to the weights of the InSAR monitoring result and the UAV-LiDAR monitoring result, the weighted fusion settlement value is obtained by using the following formula:

[0068] D′(x i , y i ) = W InSAR D InSAR (x i , y i ) + W UAV D UAV (x i , y i )

[0069] Among them, (x i , y i ) represents the coordinates of the i-th point, D InSAR (x i , y i ) represents the settlement value observed by SBAS-InSAR at a certain point, D UAV (x i , y i ) represents the settlement value measured by UAV-LiDAR at the i-th point, and D'(x i , y i ) represents the weighted fusion settlement value at the i-th point;

[0070] Exemplarily, the method provided by this application can reasonably define the transition region, and based on the idea of coherence, the heterogeneity of different data sources (SBAS-InSAR and UAV-LiDAR) can be explored, so as to characterize the relationship between point deformation and its coherence, and combined with the above data fusion strategy and model, the refined settlement of the goaf transition region can be obtained, and finally a complete refined deformation field can be extracted.

[0071] Although the weighted fusion method based on coherence weight realizes data fusion in the transition region, there are still certain errors in the local region, which are mainly affected by factors such as data noise and terrain changes. Therefore, in order to further optimize the fused settlement field and improve the monitoring accuracy of the transition region, this application optimizes based on the weighted fusion settlement value D'(x i , y i ) by using a regression model to reduce local errors, make the settlement basin smoother and more reasonable, and then obtain more complete settlement information.

[0072] Assume that there are n discrete points distributed in the transition zone. Then, the final fusion settlement value D' i , y i ) of each point P(x fin (x i , y i ) can be estimated by regression optimization using the settlement value D'(x i , y i ) after weighted fusion. The mathematical expression is as follows:

[0073] D' fin (x i , y i ) = β0(x i , y i ) + β1(x i , y i )D'(x i , y i ) + ε i

[0074] where D′ fin (x i , y i ) is the final fusion settlement value corresponding to the point (x i , y i ), β0(x i , y i ) and β1(x i , y i ) are the regression coefficients corresponding to point i, ε i is the error term, and i = 1, 2, …, n, where n is the number of discrete points.

[0075] For ease of description, the variables can be denoted in the following form:

[0076]

[0077] D′ fin = Xβ + ε i

[0078] where D' fin is the deformation vector after regression optimization; X is the explanatory variable matrix, which includes the settlement value D'(x, y) after weighted fusion; β is the regression coefficient matrix; and ε i is the error vector.

[0079] Assume that the error term ε follows a normal distribution. The expression for estimating the regression coefficients using the ordinary least squares (OLS) method is as follows:

[0080]

[0081] Exemplarily, based on the SBAS-InSAR technology to process Sentinel-1 data, coherence data was obtained, and the influence boundary of mining subsidence of 10 mm was used as a limiting condition to conduct spatial coherence statistical analysis on the entire subsidence field. The statistical results show that the minimum coherence coefficient in the study area is 0.312, and the maximum coherence coefficient is 0.851. According to the ground subsidence measurement specification, SAR data with a coherence coefficient greater than 0.3 is considered reliable. Therefore, the data obtained in this application has a high credibility.

[0082] Furthermore, based on the upper and lower bounds of the transition region obtained above, where the Sentinel-1 radar wavelength λ is taken as 5.6 cm and the pixel size μ can be taken as 20 m. After calculation, the upper and lower limits of the subsidence field after coherence weight adjustment can be -101 mm and -352 mm, corresponding to the maximum coherence coefficient of 0.851 and the average coherence coefficient of 0.312. According to this standard, the study area can be divided into a high coherence region, a low coherence region, and a transition region, providing a basis for the weighted fusion of the subsequent subsidence field.

[0083] S5: Linearly superimpose the final deformation values of the high coherence region, the final deformation values of the low coherence region, and the subsidence information after regression optimization in the transition region to obtain the final subsidence field of the subsidence basin.

[0084] Exemplarily, based on the coherence region division and data fusion processing according to the coherence levels of different regions, it can be specifically divided into the following three parts:

[0085] 1. High coherence region (d ≥ d upper ): This region has a high coherence, and the InSAR technology can stably monitor the surface deformation. Therefore, the settlement value D InSAR (x i , y i ) observed by SBAS-InSAR is directly used as the final deformation value;

[0086] 2. Low coherence region (d ≤ d lower ): This region exceeds the maximum detectable deformation gradient range of InSAR and has a low coherence. There may be large errors in the SBAS-InSAR data. Therefore, in this region, the settlement value D UAV (x i , y i ) measured by UAV-LiDAR is directly used;

[0087] 3. Transition region (d lower < d < d upper):This area lies between the high coherence and low coherence regions. Both InSAR data and UAV-LiDAR data can be utilized. In this study, a weighted fusion method is adopted for data integration, as shown in the following formula.

[0088]

[0089] Among them, d is the ground settlement value to be estimated, D' fin (x i ,y i ) is the final fused settlement value at point (x i ,y i ). i = 1, 2, …, n, where n is the number of discrete points, and it can be denoted as the fusion function f(D InSAR (x i ,y i ), D UAV (x i ,y i )); D InSAR (x i ,y i ) represents the settlement value observed by SBAS-InSAR at point i, and D UAV (x i ,y i ) represents the settlement value measured by UAV-LiDAR at point i.

[0090] Exemplarily, based on the coherence data processed by the SBAS-InSAR technology, dynamic weighting and other processing are performed on the transition region. Finally, a fused ground subsidence basin deformation field is obtained, as Figure 4 shown.

[0091] Regarding the non-uniform settlement characteristics of the subsidence basin, this application adopts a zoning weighted accuracy assessment method. By dividing the subsidence basin into different regions, the accuracy of each region is evaluated separately, and the overall accuracy of the subsidence basin is calculated by combining the regional area weights.

[0092] As Figure 5 shown, this application mainly divides the ground subsidence basin into 3 regions. The central region of the subsidence basin (subsidence greater than 0.4 m): The subsidence amplitude is relatively large, and the monitoring accuracy requirement is relatively low; the fusion region (subsidence between 0.1 and 0.4 m): It is located between the subsidence center and the boundary, and the monitoring accuracy needs to balance stability and accuracy; the boundary region of the subsidence basin (subsidence less than 0.1 m): The subsidence amplitude is relatively small, and the monitoring accuracy requirement is the highest to ensure the accuracy of the boundary settlement information.

[0093] Exemplarily, the regional area weight allocation method is adopted, that is, according to the proportion of the area of each region in the area of the entire subsidence basin, the corresponding weights are calculated. The weight calculation formula is as follows:

[0094]

[0095] Among them, P i is the weight of each area in the surface subsidence basin; Area all is the total area of the entire subsidence basin, and Area i is the area of each region.

[0096] On this basis, the overall monitoring accuracy of the subsidence basin is calculated using the error propagation law:

[0097]

[0098] Among them, m1 is the mean square error of subsidence in the central area of the surface subsidence basin, m2 is the mean square error of subsidence in the fusion transition area, and m3 is the mean square error of subsidence in the boundary area of the surface subsidence basin; m z is the overall accuracy of the surface subsidence basin.

[0099] To verify the reliability and accuracy of this method, 47 leveling points were selected as control data in this paper for comparative analysis of the results of the fused full subsidence basin and the single UAV-LiDAR and InSAR monitoring. Specifically, the results of the combination of these three different monitoring technologies were selected and compared with the measured data of the leveling points to evaluate their monitoring accuracy in different subsidence areas.

[0100] First, according to the above model, the areas of each region were statistically analyzed, and their respective weights were calculated. As shown in Table 1, the weights of each region were determined to be 0.47, 0.12, and 0.41. Secondly, according to the point distributions of the 47 leveling points, they were divided into three regions of the subarea weighted accuracy evaluation model. Among them, the central area of the subsidence basin with a subsidence value greater than 0.4 m contains 24 leveling points; the fusion area of the subsidence basin with a subsidence value between 0.1 m and 0.4 m contains 7 leveling points; the boundary area of the subsidence basin with a subsidence value less than 0.1 m contains 16 leveling points. Finally, the mean square error of the fused basin compared with the basins obtained by the two methods of single monitoring in the three regions and the overall mean square error were calculated respectively.

[0101] Table 1 Results of weight determination for accuracy evaluation

[0102]

[0103] Table 2 Accuracy statistical results of 3 methods

[0104]

[0105] From Figure 6It can be seen that, excluding some outlier points, the settlement values of the observation points extracted from the fused basin are basically consistent with those of the leveling points. The accuracy of the fused basin in the boundary area and the central area of the settlement basin has been significantly improved compared with the single UAV-LiDAR and InSAR monitoring methods.

[0106] Table 2 shows the accuracy calculation results of the surface subsidence basins obtained by the UAV-LiDAR and InSAR technologies respectively using the zonal weighted accuracy evaluation model in 3 regions and the overall accuracy. It can be seen that the accuracy of the single UAV-LiDAR monitoring is relatively high in the central area and even the fusion area of the basin, while the accuracy in the boundary area is relatively low; the accuracy of the InSAR technology monitoring is high in the boundary area of the subsidence basin. In the other two regions, due to large-scale subsidence (such as the maximum subsidence of the working face being 5.41 m), the mean square error in the fusion area and the boundary area is the lowest, and the mean square error in the central area even reaches 4293.7 mm; the overall mean square error of the fused basin has been significantly improved compared with the separate monitoring of the subsidence basin by LiDAR or InSAR. Compared with the UAV-LiDAR monitoring of the subsidence basin, the overall RMSE accuracy of the fused basin has increased by 28.4%. For the InSAR monitoring of the subsidence basin, the overall RMSE accuracy of the fused basin has increased by 98.4%.

[0107] The above is the surface subsidence monitoring method based on data fusion provided by one or more embodiments of this specification. Based on the same idea, this specification also provides a corresponding surface subsidence monitoring device based on data fusion, including:

[0108] An acquisition module, configured to acquire the settlement values observed by UAV-LiDAR and the settlement values observed by SBAS-InSAR in the same time dimension in the settlement area;

[0109] A division module, configured to obtain the coherence coefficient in the settlement area according to the acquired SBAS-InSAR data, divide the settlement area based on the coherence coefficient to obtain a high coherence area, a low coherence area, and a transition area;

[0110] A determination module, configured to use the settlement value observed by SBAS-InSAR as the final deformation value of the high coherence area, and use the settlement value measured by UAV-LiDAR as the final deformation value of the low coherence area; determine the corresponding weight according to the coherence coefficient value of each point in the transition area, fuse the settlement values observed by UAV-LiDAR and SBAS-InSAR for each point in the transition area to obtain the weighted fused settlement value, and according to the weighted fused settlement value, combine the regression coefficient determined by the least squares method OLS to optimize the weighted fused settlement value through the regression coefficient to obtain the settlement information after regression optimization in the transition area;

[0111] A fusion module is used to linearly superimpose the final deformation values in the high coherence region, the final deformation values in the low coherence region, and the settlement information after regression optimization in the transition region to obtain the final settlement field of the subsidence basin.

[0112] For the specific limitations of the ground settlement monitoring device based on data fusion, reference can be made to the limitations of the ground settlement monitoring method based on data fusion in the above text, which will not be elaborated here. Each module in the above ground settlement monitoring device based on data fusion can be implemented in whole or in part by software, hardware, and their combinations. The above modules can be embedded in the processor in the computer device in the form of hardware or independent of it, or stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0113] The present invention also provides Figure 7 a schematic structural diagram of the computer device as shown in Figure 7 As shown, at the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Of course, there may also be other hardware required for other services. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to implement the ground settlement monitoring method based on data fusion provided in the above embodiments.

[0114] Those of ordinary skill in the art can understand that all or part of the processes in the above method embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above method embodiments. Among them, any reference to the memory, storage, database, or other media used in the various embodiments provided by the present invention can include at least one of non-volatile and volatile memories. The non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical memory, etc. The volatile memory can include random access memory (RAM) or an external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0115] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as these combinations of technical features do not conflict, they should be considered as falling within the scope recorded by the present invention.

Claims

1. A ground settlement monitoring method based on data fusion, characterized in that, Including: Obtain the settlement values observed by UAV-LiDAR and the settlement values observed by SBAS-InSAR within the settlement area in the same time dimension; Based on the obtained SBAS-InSAR data, obtain the coherence coefficient within the settlement area, and divide the settlement area based on the coherence coefficient to obtain a high coherence area, a low coherence area, and a transition area; Use the settlement value observed by SBAS-InSAR as the final deformation value of the high coherence area, and use the settlement value measured by UAV-LiDAR as the final deformation value of the low coherence area; According to the coherence coefficient value of each point within the transition area, determine the corresponding weight, fuse the settlement values observed by UAV-LiDAR and SBAS-InSAR for each point within the transition area to obtain the weighted fused settlement value, and based on the weighted fused settlement value, combine the regression coefficient determined by the least squares method OLS, and optimize the weighted fused settlement value through the regression coefficient to obtain the settlement information after regression optimization in the transition area; Linearly superimpose the final deformation value of the high coherence area, the final deformation value of the low coherence area, and the settlement information after regression optimization in the transition area to obtain the final settlement field of the subsidence basin.

2. The surface subsidence monitoring method based on data fusion according to claim 1, wherein The division of the settlement area based on the coherence coefficient specifically includes: According to the coherence coefficient, use the following formula to obtain the maximum detectable deformation gradient: where d is the maximum detectable deformation gradient, γ is the coherence coefficient, and λ and μ are the radar wavelength and the pixel or spatial resolution of a single SAR image respectively; Divide the settlement area according to the relationship between the coherence coefficient and the detectable deformation gradient.

3. The surface subsidence monitoring method based on data fusion according to claim 1, characterized in that The step of determining the corresponding weight according to the coherence coefficient value of each point within the transition area, fusing the settlement values observed by UAV-LiDAR and SBAS-InSAR for each point within the transition area to obtain the weighted fused settlement value, and based on the weighted fused settlement value, combining the regression coefficient determined by the least squares method OLS, and optimizing the weighted fused settlement value through the regression coefficient to obtain the settlement information after regression optimization in the transition area specifically includes: According to the coherence coefficient value of each point within the transition area, use the following formula to obtain the weight of the InSAR monitoring result and the weight of the UAV-LiDAR monitoring result: W UAV = 1 - W InSAR Among them, r ave represents the average coherence value in the transition region, and r i is the coherence value at the i-th point in the transition region. W InSAR represents the weight of the InSAR monitoring result, and W UAV represents the weight of the UAV-LiDAR monitoring result; According to the weight of the InSAR monitoring result and the weight of the UAV-LiDAR monitoring result, use the following formula to obtain the weighted fused settlement value: D′(x i ,y i ) = W InSAR D InSAR (x i ,y i ) + W UAV D UAV (x i ,y i ) Among them, (x i , y i ) represents the coordinates of the i-th point, D InSAR (x i , y i ) represents the settlement value of the i-th point observed by SBAS-InSAR, D UAV (x i , y i ) represents the settlement value of the i-th point measured by UAV-LiDAR, D'(x i , y i ) represents the settlement value of the i-th point after weighted fusion; Based on the weighted fused settlement value, combine the regression coefficient determined by the least squares method OLS, and perform regression optimization estimation on the weighted fused settlement value through the regression coefficient. Use the following formula to obtain the settlement information after regression optimization in the transition area: D' fin (x i ,y i ) = β0(x i ,y i ) + β1(x i ,y i )D'(x i ,y i ) + ε i Among them, D' fin (x i , y i ) is the final fusion settlement value of the transition region, β0(x i , y i ) and β1(x i , y i ) are both the regression coefficients corresponding to the i-th point, ε i is the error term, i = 1, 2, …, n, and n is the number of discrete points.

4. The ground settlement monitoring method based on data fusion according to claim 1, characterized in that The step of linearly superimposing the final deformation value of the high coherence area, the final deformation value of the low coherence area, and the settlement information after regression optimization in the transition area to obtain the final settlement field of the subsidence basin specifically includes: Use the following formula to obtain the ground settlement value to be estimated: where d is the ground settlement value to be estimated, D' fin (x i , y i ) is the settlement information after the regression optimization of the transition area. i = 1, 2, …, n, where n is the number of discrete points, D InSAR (x i , y i ) represents the settlement value observed by SBAS-InSAR at the i-th point, and D UAV (x i , y i ) represents the settlement value measured by UAV-LiDAR at the i-th point.

5. The ground settlement monitoring method based on data fusion according to claim 1, characterized in that The obtaining of the settlement values observed by UAV-LiDAR and the settlement values observed by SBAS-InSAR in the same time dimension within the settlement area specifically includes: Resampling and coordinate transformation are performed on the settlement values observed by UAV-LiDAR so that the settlement values observed by UAV-LiDAR are the same as the settlement values observed by SBAS-InSAR in terms of spatio-temporal resolution and coordinate system.

6. A ground settlement monitoring device based on data fusion, characterized in that, It includes: An acquisition module for acquiring the settlement values observed by UAV-LiDAR and the settlement values observed by SBAS-InSAR in the same time dimension within the settlement area; A division module for obtaining the coherence coefficient within the settlement area based on the acquired SBAS-InSAR data, and dividing the settlement area based on the coherence coefficient to obtain a high coherence area, a low coherence area, and a transition area; A determination module for using the settlement value observed by SBAS-InSAR as the final deformation value of the high coherence area and the settlement value measured by UAV-LiDAR as the final deformation value of the low coherence area; According to the coherence coefficient value of each point in the transition area, the corresponding weight is determined, and the settlement values observed by UAV-LiDAR and SBAS-InSAR for each point in the transition area are fused to obtain the weighted fused settlement value. According to the weighted fused settlement value, combined with the regression coefficient determined by the least squares method OLS, the weighted fused settlement value is optimized through the regression coefficient to obtain the settlement information after regression optimization in the transition area; A fusion module for linearly superimposing the final deformation value of the high coherence area, the final deformation value of the low coherence area, and the settlement information after regression optimization in the transition area to obtain the final subsidence basin settlement field.

7. A computer device, characterized in that, It includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the surface settlement monitoring method based on data fusion according to any one of claims 1 to 5 above.

Images