Landslide monitoring method and device

Abstract

The present invention provides a landslide monitoring method and device, the method comprising: obtaining multi-source data to be processed; obtaining a first feature based on optical remote sensing data, obtaining a second feature based on radar remote sensing data, and obtaining a third feature based on multi-source data; performing spatiotemporal analysis on the first feature, the second feature, and the third feature, respectively, to obtain slope creep information. The method of the present invention extracts three features from the multi-source data of the slope, and combines the temporal and spatial changes of the three features to perform a comprehensive analysis of the slope landslide situation, thereby achieving effective monitoring of slope creep.

Description

Technical Field

[0001] The invention relates to the technical field of remote sensing image processing, and in particular to a landslide monitoring method and device. Background Art

[0002] Landslide is a common geological disaster, posing a great threat to people's lives and travel.

[0003] In related technologies, the displacement of slopes is usually monitored using methods such as the Global Navigation Satellite System (GNSS) and Interferometric Synthetic Aperture Radar (InSAR). However, when monitoring landslides in complex terrain with low coherent point density and uneven spatial distribution, landslide monitoring equipment based on GNSS and InSAR technology is difficult to deploy, and the extracted slope displacement data has low accuracy, resulting in a decrease in the accuracy of monitoring slope creep. Summary of the invention

[0004] The present invention provides a landslide monitoring method and device, which are used to solve the defects of the prior art that when monitoring landslides using GNSS and InSAR technologies, the landslides are easily affected by terrain factors such as low density of coherent points and uneven spatial distribution, resulting in difficulty in arranging landslide monitoring equipment and low accuracy of extracted slope displacement data, thereby improving the accuracy of monitoring slope creep.

[0005] The present invention provides a landslide monitoring method, comprising:

[0006] Acquire multi-source data to be processed, wherein the multi-source data includes optical remote sensing data of a slope, radar remote sensing data of the slope, and digital elevation model (DEM) data of the slope;

[0007] Based on the optical remote sensing data, a first feature is obtained, based on the radar remote sensing data, a second feature is obtained, and based on the multi-source data, a third feature is obtained, wherein the first feature is used to represent the change in vegetation coverage of the slope, the second feature is used to represent the deformation information of the slope, and the third feature is used to represent the landslide sensitivity of the slope;

[0008] The first feature, the second feature and the third feature are respectively subjected to spatiotemporal analysis to obtain creep information of the slope.

[0009] According to a landslide monitoring method provided by the present invention, the first feature, the second feature and the third feature each include a plurality of first features, the plurality of first features are obtained based on optical remote sensing data corresponding to the plurality of regions of the slope in the first time unit, the plurality of second features are obtained based on radar remote sensing data corresponding to the plurality of regions of the slope in the first time unit, and the plurality of third features are obtained based on multi-source data corresponding to the plurality of regions of the slope in the first time unit, and the first feature, the second feature and the third feature are respectively subjected to spatiotemporal analysis to obtain deformation information of the slope, including:

[0010] The spatial creep information of the slope is obtained by comparing the plurality of the first features, comparing the plurality of the second features, and comparing the plurality of the third features.

[0011] According to a landslide monitoring method provided by the present invention, the first feature, the second feature and the third feature each include a plurality of first features, the plurality of first features are obtained based on optical remote sensing data corresponding to the first area of ​​the slope in a plurality of time units, the plurality of second features are obtained based on radar remote sensing data corresponding to the first area of ​​the slope in a plurality of time units, and the plurality of third features are obtained based on multi-source data corresponding to the first area of ​​the slope in a plurality of time units, and the first feature, the second feature and the third feature are respectively subjected to spatiotemporal analysis to obtain deformation information of the slope, further comprising:

[0012] The time creep information of the slope is obtained by comparing the plurality of the first features, comparing the plurality of the second features, and comparing the plurality of the third features.

[0013] According to a landslide monitoring method provided by the present invention, obtaining a third feature based on the multi-source data includes:

[0014] The multi-source data are input into a first model to obtain a third feature output by the first model, wherein the first model is obtained based on a landslide sensitivity factor and a weight of the landslide sensitivity factor.

[0015] According to a landslide monitoring method provided by the present invention, the landslide sensitivity factors include multiple factors, and the first model is obtained by the following steps:

[0016] constructing a first matrix based on the plurality of landslide sensitivity factors, wherein the first matrix is ​​used to determine the weights of the landslide sensitivity factors;

[0017] Based on the first matrix, obtaining a normalized weight of the landslide sensitivity factor;

[0018] The first model LSI is obtained by the following formula:

[0019] LSI=∑wi ×f i

[0020] Among them, f i represents the ith sensitivity factor, w i represents the normalized weight corresponding to the i-th landslide sensitivity factor, i is a natural number greater than 0, 0 <w i <1.

[0021] According to a landslide monitoring method provided by the present invention, obtaining a first feature based on the optical remote sensing data includes:

[0022] The normalized difference vegetation index NDVI corresponding to the optical remote sensing data is obtained by the following formula:

[0023]

[0024] Among them, ρ NIR represents the reflectivity of the near-infrared band in the optical remote sensing data, ρ R represents the reflectivity of the red band in the optical remote sensing data;

[0025] Performing partition processing on the optical remote sensing data to obtain a plurality of partition data;

[0026] The first features corresponding to the plurality of partition data are obtained based on a pixel binary model, wherein the pixel binary model is determined by the following formula:

[0027]

[0028] Among them, NDVI soil Indicates the normalized difference vegetation index (NDVI) of pure vegetation corresponding to the partition data veg The normalized difference vegetation index of pure bare land corresponding to the partition data.

[0029] According to a landslide monitoring method provided by the present invention, the second feature includes the surface deformation amount of the slope and the surface deformation rate of the slope, and the second feature is obtained based on the radar remote sensing data, including:

[0030] Extracting a time baseline and a space baseline from the radar remote sensing data based on the short baseline synthetic aperture radar interferometry (SBAS-InSAR) technology;

[0031] Based on the time baseline and the space baseline, a connection graph of the radar remote sensing data is obtained;

[0032] Performing interference measurement on the connection diagram to obtain a differential interference diagram;

[0033] The differential interference pattern is subjected to phase unwrapping, singular value decomposition, and atmospheric phase removal operations to obtain the surface deformation amount and the surface deformation rate.

[0034] The present invention provides a landslide monitoring device, comprising:

[0035] An acquisition module is used to acquire multi-source data to be processed, wherein the multi-source data includes optical remote sensing data of the slope, radar remote sensing data of the slope, and digital elevation model DEM data of the slope;

[0036] A feature extraction module, used to obtain a first feature based on the optical remote sensing data, a second feature based on the radar remote sensing data, and a third feature based on the multi-source data, wherein the first feature is used to represent the change in vegetation coverage of the slope, the second feature is used to represent the deformation information of the slope, and the third feature is used to represent the landslide sensitivity of the slope;

[0037] The analysis module is used to perform spatiotemporal analysis on the first feature, the second feature and the third feature respectively to obtain creep information of the slope.

[0038] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the landslide monitoring methods described above is implemented.

[0039] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the landslide monitoring method described in any one of the above methods is implemented.

[0040] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the landslide monitoring method described above is implemented.

[0041] The landslide monitoring method and device provided by the present invention obtain multi-source data including optical remote sensing data, radar remote sensing data and digital elevation model DEM data of the slope, and respectively extract characteristic information representing the change in vegetation coverage of the slope, deformation information of the slope and landslide sensitivity of the slope from the multi-source data, and finally perform time analysis and space analysis on the characteristic information to obtain creep information of the slope; the method extracts three characteristics from the multi-source data of the slope, and comprehensively analyzes the landslide situation of the slope by combining the time change and space change of the three characteristics, thereby realizing effective monitoring of slope creep. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are 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.

[0043] Figure 1 It is one of the flow diagrams of the landslide monitoring method provided by the present invention;

[0044] Figure 2 This is the second flow chart of the landslide monitoring method provided by the present invention;

[0045] Figure 3 It is a schematic diagram of the numerical changes of the first characteristic, the second characteristic and the third characteristic provided by the present invention in different cross sections of the slope;

[0046] Figure 4 It is a schematic diagram of numerical changes of the first feature, the second feature and the third feature provided by the present invention in different partitions of the slope;

[0047] Figure 5 It is a schematic diagram of the process flow of the optical remote sensing image preprocessing method provided by the present invention;

[0048] Figure 6 It is a structural schematic diagram of the landslide monitoring device provided by the present invention;

[0049] Figure 7 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] Combine the following Figure 1-Figure 6 The landslide monitoring method and device of the present invention are described.

[0052] like Figure 1 As shown, the landslide monitoring method includes:

[0053] Step 110: Acquire multi-source data to be processed, where the multi-source data includes optical remote sensing data of the slope, radar remote sensing data of the slope, and digital elevation model (DEM) data of the slope.

[0054] It should be noted that the multi-source data can be a variety of remote sensing data or elevation data of the slope before the landslide, which are used to represent the vegetation coverage, deformation and landslide sensitivity of the slope before the landslide. Combining these characteristic information can be used to predict the probability of a landslide before the slope landslide; the multi-source data can be a variety of remote sensing data or elevation data of the slope after the landslide occurs. Combining the multi-source data can be used to predict the probability of a landslide again after the landslide.

[0055] In this step, the multi-source data is a collection of multiple remote sensing data or DEM data collected based on the area where the slope is located, wherein the optical remote sensing data can be time-series optical remote sensing high-resolution data of the slope, and the radar remote sensing data can be time-series SAR remote sensing data.

[0056] In this embodiment, the optical remote sensing data of the slope can adopt the Gaofen-1 high-resolution image data in the remote sensing data set or database, the radar remote sensing data of the slope can adopt the Sentinel-1A data, and the digital elevation model data can select the ALOS-PALSAR (Advanced Land Observation Satellite) DEM data.

[0057] In some embodiments, the multi-source data may include remote sensing data or DEM data of the slope, and may also include remote sensing data or DEM data of the slope and its surrounding areas.

[0058] Step 120: obtain a first feature based on optical remote sensing data, obtain a second feature based on radar remote sensing data, and obtain a third feature based on multi-source data, wherein the first feature is used to represent the change in vegetation coverage of the slope, the second feature is used to represent the deformation information of the slope, and the third feature is used to represent the landslide sensitivity of the slope.

[0059] It should be noted that vegetation, as the covering layer of the slope surface, can reflect the internal activities of the slope and the influence of various factors. The vegetation coverage (Fractional Vegetation Cover, FVC) is defined as the percentage of the vegetation projection area to the statistical area, which is an indicator used to directly characterize the vegetation growth status. The pixel binary model divides the information obtained by the sensor into two parts: vegetation and soil, and the sum of these two parts in the pixel area is 1.

[0060] In this step, the first feature can reflect the growth of vegetation on the slope, and then reflect the activities inside the slope. For example, vegetation coverage is related to the creep of landslides and the distribution of landslide points. The larger the NDVI value and the larger the FVC value, the smaller the probability of landslide.

[0061] In this embodiment, the Normalized Difference Vegetation Index (NDVI) is calculated based on the vegetation distribution on the slope in the Gaofen-1 data, and a confidence interval is set to count the FVC value on the landslide, and the cumulative percentages of the NDVI values ​​corresponding to when the slope surface is pure vegetation and pure bare land, respectively, are used as the upper and lower limits of the confidence interval.

[0062] In this step, the second feature is used to represent the deformation amount and deformation rate of the slope. The larger the deformation amount and the larger the deformation rate of the slope, the greater the probability of landslide on the slope.

[0063] In this embodiment, the Sentinel-1A data is input into ENVI SARScape for SBAS-InSAR processing to obtain the second feature of the slope.

[0064] In some embodiments, the cumulative deformation map and deformation rate map extracted based on the second feature can be imported into the map service software to obtain the geographical location information of the slope containing the second feature, so as to check the deformation rate of the slope.

[0065] In this step, the third feature represents the probability, susceptibility or danger of landslides in a specific area. The third feature is negatively correlated with the first feature, that is, the larger the FVC value, the lower the landslide sensitivity of the slope. The third feature is negatively correlated with the second feature, and the value of the second feature is less than 0, that is, the smaller the second feature of the slope, the larger the corresponding actual deformation, and the higher the landslide sensitivity.

[0066] In this embodiment, there are many factors that affect the landslide susceptibility of the slope, such as the elevation, slope, terrain undulation, terrain moisture index TWI (Topographic Wetness Index), NDVI and landslide deformation of the slope. These factors can be extracted from the optical remote sensing data, radar remote sensing data and DEM data of the slope respectively.

[0067] In this embodiment, a landslide sensitivity index model can be constructed based on the above-mentioned landslide sensitivity factors and the hierarchical analysis method. The sensitivity index model is determined by the above-mentioned factors and their corresponding weights. The landslide sensitivity of the slope is calculated year by year in the model, so as to subsequently perform temporal and spatial analysis on the landslide sensitivity of the slope.

[0068] In this embodiment, various landslide sensitivity factors of the slope are extracted from the Gaofen-1 data, Sentinel-1A data, and DEM data. After determining the weights corresponding to each factor, a sensitivity index model can be obtained to determine the landslide sensitivity of the slope year by year.

[0069] Step 130: Perform spatiotemporal analysis on the first feature, the second feature, and the third feature to obtain creep information of the slope.

[0070] In this step, spatiotemporal analysis includes time analysis and space analysis.

[0071] In this step, a spatial analysis is performed on the first feature, which can be to analyze the pixel changes of higher vegetation coverage (50% ≤ FVC < 70%) or high vegetation coverage (FVC ≥ 70%) in different spaces of the landslide (along the direction of landslide movement or different sub-areas), and analyze the reasons; a temporal analysis is performed on the first feature, which is to analyze the temporal changes of vegetation coverage in a fixed area or point of the slope at the same spatial position of the slope and at different times.

[0072] In this step, a spatial analysis is performed on the second feature, which can be to analyze the changes in the surface deformation amount or deformation rate of the slope in different spaces of the landslide (along the direction of landslide movement or different sub-areas), and analyze the reasons; a temporal analysis is performed on the second feature, which can be to analyze the time series change characteristics of the surface deformation amount or deformation rate of the slope at the same spatial position of the slope and at different times.

[0073] In this step, a spatial analysis is performed on the third feature, which can be to analyze the changes in landslide sensitivity in different spaces of the landslide (along the direction of landslide movement or in different sub-areas), and analyze its causes; a temporal analysis is performed on the second feature, which can be to analyze the temporal changes in the landslide sensitivity of the slope at the same spatial position of the slope and at different times.

[0074] In some embodiments, the probability of slope creep or the degree of slope deformation can be determined based on the analysis results of time analysis and space analysis of the first feature, the second feature, and the third feature.

[0075] The landslide monitoring method provided by the present invention obtains multi-source data including optical remote sensing data, radar remote sensing data and digital elevation model DEM data of the slope, and extracts characteristic information representing the change in slope vegetation coverage, slope deformation information and landslide sensitivity of the slope from the multi-source data, and finally performs time analysis and spatial analysis on the characteristic information to conduct a comprehensive analysis of the slope landslide situation, thereby realizing effective monitoring of slope creep.

[0076] In some embodiments, the first feature, the second feature and the third feature each include multiple first features, the multiple first features are obtained based on optical remote sensing data corresponding to multiple regions of the slope in the first time unit, the multiple second features are obtained based on radar remote sensing data corresponding to multiple regions of the slope in the first time unit, and the multiple third features are obtained based on multi-source data corresponding to multiple regions of the slope in the first time unit. The first feature, the second feature and the third feature are respectively subjected to spatiotemporal analysis to obtain deformation information of the slope, including: comparing the multiple first features, comparing the multiple second features, and comparing the multiple third features to obtain spatial creep information of the slope.

[0077] In this embodiment, the second characteristic may include the deformation rate and deformation amount of the slope.

[0078] In this embodiment, the variation patterns of vegetation coverage, deformation, and landslide sensitivity on the whole landslide are first considered.

[0079] In this embodiment, after spatial analysis of the first feature, the second feature and the third feature, it is found that the deformation variables of the first feature and the second feature are positively correlated in the change of the spatial position of the slope, and the deformation variables of the first feature and the second feature are negatively correlated with the change of the spatial position of the third feature in the slope.

[0080] In this embodiment, since SBAS-InSAR obtains deformation information along the satellite line of sight LOS (Line of Sight), deformation information along the landslide movement direction (from the top to the bottom of the landslide) can be obtained along the longitudinal section; along the longitudinal section of the landslide, a section line is drawn from the boundary of the rear edge of the landslide to the boundary line of the circular arc front edge of the landslide, and the changes of the three along the section line are analyzed.

[0081] exist Figure 3 In the embodiment shown, Figure 3 (a) to Figure 3 (c) The numerical variation diagrams of the first, second and third characteristics of the slope at different cross sections of the slope respectively show that when the value of the first characteristic FVC decreases, the vegetation condition deteriorates, the deformation value of the second characteristic decreases, and the deformation increases; when the third characteristic landslide sensitivity index decreases, it means that the landslide sensitivity decreases and the landslide condition improves. Therefore, the variation trend of FVC is consistent with that of deformation, and opposite to that of LSI; Figure 3 At the straight line where point x is located, when the FVC is at a small trough, the deformation variable is also at a low point, and the LSI is at a small peak. The same rule applies to the straight line where point z is located. At the straight line where point y is located, FVC is at a small peak, the deformation variable is at a higher point, and LSI is at a small low peak.

[0082] The landslide monitoring method provided by the present invention determines the spatial variation trend of each feature by analyzing the spatial variation laws of the first feature, the second feature and the third feature of the slope, which helps to comprehensively evaluate the creep condition of the slope from both time and space aspects, thereby improving the accuracy of slope creep monitoring.

[0083] In some embodiments, the first feature, the second feature and the third feature each include multiple first features, the multiple first features are obtained based on optical remote sensing data corresponding to the first area of ​​the slope in multiple time units, the multiple second features are obtained based on radar remote sensing data corresponding to the first area of ​​the slope in multiple time units, and the multiple third features are obtained based on multi-source data corresponding to the first area of ​​the slope in multiple time units. The first feature, the second feature and the third feature are respectively subjected to spatiotemporal analysis to obtain deformation information of the slope, which also includes: comparing the multiple first features, comparing the multiple second features, and comparing the multiple third features to obtain time creep information of the slope.

[0084] In this embodiment, after the spatial analysis of the first feature, the second feature and the third feature, the creep information of the slope can also be obtained according to the time variation characteristics of the above features.

[0085] In this embodiment, the deformation rates of the first feature and the second feature are positively correlated in the temporal change of the slope, and the deformation rates of the first feature and the second feature are negatively correlated with the third feature in the temporal change of the slope.

[0086] exist Figure 4 In the embodiment shown, Figure 4 (a) to Figure 4 (c) The numerical change diagrams of the first, second and third characteristics of the slope in different partitions of the slope respectively; first, three central points were selected in the three partitions selected on the slope to analyze the changes of FVC, LSI and deformation rate over time; the deformation rate has values ​​in the 62 selected time phases, and most of them are negative, indicating that the deformation is caused by the sinking movement of the landslide; in the annual changes of the values ​​of the three points, it can be clearly found that FVC and deformation rate both conform to the "increase-increase-decrease" law between 2018 and 2021, while LSI shows the opposite law, that is, "decrease-decrease-increase", which is in line with the regional average level law.

[0087] The landslide monitoring method provided by the present invention determines the time variation trend of each feature by analyzing the time variation law of the first feature, the second feature and the third feature of the slope, which helps to comprehensively evaluate the creep condition of the slope from both time and space aspects, thereby improving the accuracy of slope creep monitoring.

[0088] In some embodiments, obtaining the third feature based on multi-source data includes: inputting the multi-source data into a first model to obtain the third feature output by the first model, wherein the first model is obtained based on a landslide sensitivity factor and a weight of the landslide sensitivity factor.

[0089] In this embodiment, the multi-source data includes optical remote sensing data, radar remote sensing data and digital elevation model DEM data of the slope. For example, the multi-source data includes Gaofen-1 data, Sentinel-1A data and DEM data selected from ALOS-PALSAR (Advanced Land Observation Satellite).

[0090] In this embodiment, the landslide sensitivity factor may be the elevation of the slope, the slope, the terrain relief, the terrain moisture index, the NDVI, and the deformation of the landslide, etc. These factors are respectively determined by three types of data in the multi-source data.

[0091] In this embodiment, the multiple weights corresponding to the multiple landslide sensitivity factors may represent the influence between different factors.

[0092] In some embodiments, multiple weights may be normalized.

[0093] The landslide monitoring method provided by the present invention determines the third feature through multiple landslide sensitivity factors and corresponding weights to represent the sensitivity of slope deformation, and combines the first feature and the second feature to monitor slope deformation information, thereby improving monitoring accuracy.

[0094] In some embodiments, the landslide sensitivity factor includes multiple factors, and the first model is obtained by the following steps: constructing a first matrix based on multiple landslide sensitivity factors, the first matrix is ​​used to determine the weight of the landslide sensitivity factor; based on the first matrix, obtaining the normalized weight of the landslide sensitivity factor; obtaining the first model LSI by the following formula:

[0095] LSI=∑w i ×f i ,

[0096] Among them, f i represents the i-th sensitivity factor, w i represents the normalized weight corresponding to the i-th landslide sensitivity factor, i is a natural number greater than 0, 0 <w i <1.

[0097] In this embodiment, the landslide sensitivity factors may be the elevation of the slope, the slope, the terrain relief, the terrain moisture index, the NDVI, and the deformation of the landslide, etc. These landslide sensitivity factors may be extracted from different data in the multi-source data.

[0098] In this embodiment, the first matrix may be a pairwise judgment matrix, which includes the influence relationship between various landslide sensitivity factors and the weight of each factor; the influence relationship may be a direct influence or an indirect influence.

[0099] In some embodiments, the influence relationship between the landslide sensitivity factors can be represented by a weighted scoring method, for example, a 0.5-1 weighted scoring method, where two factors with a direct influence relationship are scored as 1, and two factors with an indirect influence relationship are scored as 0.5, thereby clearly displaying the influence relationship between the factors and the weight of each factor in the first matrix.

[0100] In this embodiment, the landslide sensitivity factors include the elevation of the slope, the slope, the terrain undulation, the terrain moisture index, the NDVI and the deformation of the landslide, wherein the judgment weight is the sum of the weights of each factor and other factors scored using the 0.5-1 weight scoring method.

[0101] In this embodiment, the judgment weights of the landslide sensitivity factors are normalized to obtain corresponding normalized weights, and the first model LSI of the slope is obtained based on the normalized weights.

[0102] In Table 1, the normalized weights corresponding to elevation (A), slope (B), terrain relief (C), terrain wetness index (D), NDVI (E) and landslide deformation (F) are 0.16, 0.2, 0.18, 0.14, 0.16 and 0.16 respectively, so the first model LSI can be expressed as:

[0103] LSI = 0.16*elevation + 0.2*slope + 0.18*terrain relief + 0.14*terrain moisture index + 0.16*NDVI + 0.16*deformation.

[0104] Table 1: First matrix

[0105]

[0106] In some embodiments, the landslide sensitivity factor can be extracted from different data in the multi-source data. Before constructing the first model, the multi-source data needs to be unified in terms of coordinate system, number of rows and columns, and spatial resolution. At the same time, in order to prevent subsequent calculations from being affected by the magnitude and dimension of different factors and to ensure that each pixel in the multi-source data has a value, the landslide sensitivity factor can be normalized.

[0107] In some embodiments, after constructing the first model, it is necessary to check the consistency of the first model and calculate its consistency ratio (CR) value. When the CR value is less than the consistency threshold, it means that the first model has good consistency, that is, the first model can be used in subsequent calculation processes.

[0108] In this embodiment, the consistency threshold can be customized according to actual needs. For example, the consistency threshold can be 0.1.

[0109] The landslide monitoring method provided by the present invention realizes non-displacement monitoring of the slope by extracting characteristic factors for characterizing the landslide sensitivity of the slope from multi-source data and constructing a first model to obtain the landslide sensitivity of the slope, which helps to combine with the displacement monitoring method of the slope, thereby improving the accuracy of slope creep monitoring.

[0110] In some embodiments, obtaining the first feature based on the optical remote sensing data includes: obtaining the normalized vegetation index NDVI corresponding to the optical remote sensing data by the following formula:

[0111]

[0112] Among them, ρ NIR represents the reflectivity of the near-infrared band in optical remote sensing data, ρ R Represents the reflectivity of the red band in the optical remote sensing data; the optical remote sensing data is partitioned to obtain multiple partition data; the first feature corresponding to the multiple partition data is obtained through the pixel binary model, and the pixel binary model is determined by the following formula:

[0113]

[0114] Among them, NDVI soil Indicates the normalized difference vegetation index (NDVI) of pure vegetation corresponding to the partition data veg Represents the normalized difference vegetation index of pure bare land corresponding to the partition data.

[0115] In this embodiment, the vegetation coverage is obtained through a pixel binary model. The pixel binary model divides the information obtained by the sensor into two parts: vegetation and soil. That is, the sum of the vegetation and soil in the pixel area composition is 1.

[0116] In this embodiment, pure vegetation and pure bare land can be clearly found through optical remote sensing data, but due to differences in environment and time, the NDVI corresponding to vegetation and bare land is soil and NDVI veg The values ​​are not consistent. When using the pixel dichotomy model to determine NDVI soil and NDVI vegWhen NDVI is used, you can set a confidence interval in the NDVI statistics table, and the upper limit of the interval corresponds to the NDVI veg , the lower limit of the interval corresponds to NDVI soil The NDVI statistical table is obtained by collecting NDVI data from multiple partitions of the slope body on the optical remote sensing data. The NDVI value of each partition in the NDVI statistical table can be calculated by the following formula:

[0117]

[0118] In some embodiments, after obtaining the first characteristic FVC (vegetation coverage) of the plurality of partitions on the slope, the FVC of each partition of the slope is processed by the following steps:

[0119] (1) The pseudo-color density segmentation method is used to perform pseudo-color enhancement on FVC to obtain an FVC that is beneficial to human eye judgment.

[0120] (2) Using the vegetation coverage classification standard, the slope vegetation coverage is divided into five levels: low vegetation coverage (FVC < 10%), relatively low vegetation coverage (10% ≤ FVC < 30%), moderate vegetation coverage (30% ≤ FVC < 50%), relatively high vegetation coverage (50% ≤ FVC < 70%), and high vegetation coverage (FVC ≥ 70%).

[0121] (3) Using a higher vegetation coverage threshold (50%) or a high vegetation coverage threshold (70%), the cumulative percentage of pixels in a certain area with vegetation coverage below a certain threshold is calculated to obtain the proportion of pixels with higher vegetation coverage or high vegetation coverage.

[0122] (4) Perform temporal and spatial analysis on the first characteristic of the slope. Temporal analysis includes analyzing the temporal changes of vegetation coverage in a fixed partition or point of the slope at different times in the same space of the landslide; spatial analysis includes analyzing the changes of pixels with higher vegetation coverage or high vegetation coverage in different spaces of the landslide (along the direction of landslide movement or in different partitions), and analyzing the reasons.

[0123] The landslide monitoring method provided by the present invention obtains multiple partition data by partitioning the optical remote sensing data, and obtains the proportion of pixels with higher vegetation coverage or high vegetation coverage in the above partition based on the vegetation coverage threshold, thereby realizing non-displacement monitoring of the slope, which is helpful to combine with the displacement monitoring method of the slope, thereby improving the accuracy of slope creep monitoring.

[0124] In some embodiments, the second feature includes the surface deformation amount of the slope and the surface deformation rate of the slope. The second feature is obtained based on radar remote sensing data, including: extracting the time baseline and the space baseline from the radar remote sensing data based on the short baseline synthetic aperture radar interferometry (SBAS-InSAR) technology; obtaining a connection graph of the radar remote sensing data based on the time baseline and the space baseline; performing interferometry measurement on the connection graph to obtain a differential interferogram; performing phase unwrapping, singular value decomposition, and removal of the atmospheric phase of the interferogram on the differential interferogram to obtain the surface deformation amount and the surface deformation rate.

[0125] It should be noted that the SBAS-InSAR method can be used to extract the second feature of the slope. The SBAS-InSAR technology uses distributed scatterer points with strong coherence in a short period of time. Such scatterer points are commonly found in nature. Therefore, the SBAS-InSAR method can be used to monitor the deformation process in natural scenes.

[0126] In this embodiment, the radar remote sensing data is a SAR image covering the study area. Assuming that the number of SAR images is N+1 (N>0), these images are connected to each other through a suitable time baseline and space baseline to obtain L (L is a natural number greater than 1) differential interferograms, and L satisfies the following conditions:

[0127]

[0128] Assume t A and t B The image acquired at the moment (t A <t B ) generates the i-th (1≤i≤L) differential interference pattern. The coordinate of a certain pixel in the image pair is x in azimuth and r in range. Then the interference phase of the pixel is:

[0129]

[0130] Where i∈(1,2,…,j), j represents the total number of interference patterns, d(t B ,x,r) represents the pixel (x,r) at time t A and t B The accumulated deformation in the radar satellite line of sight (Line Of Sight, LOS), represents the interferometric residual terrain phase, represents the phase caused by atmospheric delay, Represents the phase caused by interference fringes, system thermal noise, etc.

[0131] In this embodiment, assuming that the residual terrain phase, the atmospheric phase, and the noise phase are not considered, the interference phase of the pixel can also be expressed as:

[0132]

[0133] Assume M P =[M1,M2,…,M j ] represents the master image, S P =[S1,S2,…,S j ] represents the slave image, from which the phase difference can be obtained as:

[0134]

[0135] in, Indicates the main image M p The phase at time t (t>0), Indicates that from the image S p The phase at time t), p∈(1,2,…,j), then the average phase change rate of two M*N matrices (M and N are both natural numbers greater than 0) can be expressed as:

[0136]

[0137] In this embodiment, singular value decomposition and least square method are used to solve the equation, and finally the deformation rate and the cumulative deformation amount are obtained.

[0138] In some embodiments, the surface deformation amount of the slope can be represented by the above-mentioned phase difference.

[0139] The landslide monitoring method provided by the present invention processes radar remote sensing images through SBAS-InSAR technology to obtain a second feature for representing the surface deformation amount and surface deformation rate of the slope, which helps to perform time analysis and spatial analysis on the second feature of the slope in the subsequent process, thereby improving the accuracy of slope creep monitoring.

[0140] In some embodiments, before the optical remote sensing data is partitioned, the optical remote sensing image is preprocessed by image correction, image enhancement and image registration to obtain the optical remote sensing data.

[0141] In this embodiment, image correction of the optical remote sensing image includes remote sensing image radiation correction and remote sensing image orthorectification, wherein the remote sensing image radiation correction is specifically: radiation calibration of the multispectral image is performed, and then FLAASH atmospheric correction is performed; for the panchromatic image, radiation correction is directly performed; and the remote sensing image orthorectification is specifically: using DEM data with a resolution of 30 meters or higher to orthorectify the optical image.

[0142] In this embodiment, image enhancement includes spatial-spectral fusion of remote sensing images, specifically, using the nearest neighbor fusion method (Nearest Neighbor Diffusion Pan Sharpening, NND) with good spectral fidelity to perform spatial-spectral fusion of multispectral images and panchromatic images, while ensuring that the spectral information remains unchanged as much as possible, the spatial resolution of the image is increased to the same resolution as the panchromatic image to achieve information enhancement.

[0143] It should be noted that, since remote sensing images need to take into account temporal changes, in order to make the images of each phase correspond to the same ground objects at the same position, the images of a set year are used as a reference to align the images of other phases. The set year can be one year or multiple years to ultimately determine the optical remote sensing data.

[0144] exist Figure 5 In the illustrated embodiment, optical remote sensing images are subjected to optical remote sensing image radiation correction, optical remote sensing image orthorectification, optical remote sensing image, optical remote sensing image spatial-spectral fusion, and optical remote sensing image registration, respectively, to complete the optical remote sensing image preprocessing process and obtain the required optical remote sensing data.

[0145] The landslide monitoring method provided by the present invention improves the reliability of data by performing preprocessing operations of image correction, image enhancement and image registration on optical remote sensing data, while ensuring the validity of the first feature extracted from the optical remote sensing data.

[0146] The landslide monitoring device provided by the present invention is described below. The landslide monitoring device described below and the landslide monitoring method described above can be referred to each other.

[0147] like Figure 6 As shown, the landslide monitoring device comprises:

[0148] An acquisition module 610 is used to acquire multi-source data to be processed, where the multi-source data includes optical remote sensing data of the slope, radar remote sensing data of the slope, and digital elevation model (DEM) data of the slope;

[0149] The feature extraction module 620 is used to obtain a first feature based on optical remote sensing data, a second feature based on radar remote sensing data, and a third feature based on multi-source data, wherein the first feature is used to represent the change in vegetation coverage of the slope, the second feature is used to represent the deformation information of the slope, and the third feature is used to represent the landslide sensitivity of the slope;

[0150] The analysis module 630 is used to perform spatiotemporal analysis on the first feature, the second feature and the third feature respectively to obtain creep information of the slope.

[0151] The landslide monitoring device provided by the present invention acquires multi-source data including optical remote sensing data, radar remote sensing data and digital elevation model DEM data of the slope through an acquisition module 610, and extracts characteristic information representing the change in slope vegetation coverage, the deformation of the slope and the landslide sensitivity of the slope from the multi-source data through a feature extraction module 620. Finally, the characteristic information is subjected to temporal and spatial analysis through an analysis module 630 to conduct a comprehensive analysis of the landslide situation of the slope, thereby realizing effective monitoring of slope creep.

[0152] Figure 7 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 7 As shown, the electronic device may include: a processor 710, a communication interface 720, a memory 730 and a communication bus 740, wherein the processor 710, the communication interface 720 and the memory 730 communicate with each other through the communication bus 740. The processor 710 may call the logic instructions in the memory 730 to execute the landslide monitoring method, which includes: obtaining multi-source data to be processed, the multi-source data including optical remote sensing data of the slope, radar remote sensing data of the slope and digital elevation model DEM data of the slope; obtaining a first feature based on the optical remote sensing data, obtaining a second feature based on the radar remote sensing data, and obtaining a third feature based on the multi-source data; performing spatiotemporal analysis on the first feature, the second feature and the third feature respectively to obtain creep information of the slope.

[0153] In addition, the logic instructions in the above-mentioned memory 730 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0154] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the landslide monitoring method provided by the above-mentioned methods. The method includes: obtaining multi-source data to be processed, the multi-source data including optical remote sensing data of the slope, radar remote sensing data of the slope, and digital elevation model DEM data of the slope; obtaining a first feature based on the optical remote sensing data, obtaining a second feature based on the radar remote sensing data, and obtaining a third feature based on the multi-source data; performing spatiotemporal analysis on the first feature, the second feature, and the third feature respectively to obtain creep information of the slope.

[0155] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it is implemented to execute the landslide monitoring method provided by the above-mentioned methods. The method includes: obtaining multi-source data to be processed, the multi-source data including optical remote sensing data of the slope, radar remote sensing data of the slope and digital elevation model DEM data of the slope; obtaining a first feature based on the optical remote sensing data, obtaining a second feature based on the radar remote sensing data, and obtaining a third feature based on the multi-source data; performing spatiotemporal analysis on the first feature, the second feature and the third feature respectively to obtain creep information of the slope.

[0156] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0157] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0158] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A landslide monitoring method, characterized in that: include: Acquire multi-source data to be processed, wherein the multi-source data includes optical remote sensing data of a slope, radar remote sensing data of the slope, and digital elevation model (DEM) data of the slope; Based on the optical remote sensing data, a first feature is obtained, based on the radar remote sensing data, a second feature is obtained, and based on the multi-source data, a third feature is obtained, wherein the first feature is used to represent the change in vegetation coverage of the slope, the second feature is used to represent the deformation information of the slope, and the third feature is used to represent the landslide sensitivity of the slope; The obtaining a third feature based on the multi-source data includes: Inputting the multi-source data into a first model to obtain a third feature output by the first model, wherein the first model is obtained based on a landslide sensitivity factor and a weight of the landslide sensitivity factor; The first model is a sensitivity index model, which is determined based on multiple landslide sensitivity factors and a hierarchical analysis method, wherein the multiple landslide sensitivity factors include the elevation of the slope, the slope, the terrain relief, the terrain moisture index, the NDVI and the deformation of the landslide; The first feature, the second feature and the third feature are respectively subjected to spatiotemporal analysis to obtain creep information of the slope.

2. The landslide monitoring method according to claim 1, characterized in that: The first feature, the second feature and the third feature each include a plurality of first features, the plurality of first features are obtained based on optical remote sensing data corresponding to the plurality of regions of the slope in the first time unit, the plurality of second features are obtained based on radar remote sensing data corresponding to the plurality of regions of the slope in the first time unit, and the plurality of third features are obtained based on multi-source data corresponding to the plurality of regions of the slope in the first time unit. The first feature, the second feature and the third feature are respectively subjected to spatiotemporal analysis to obtain deformation information of the slope, including: The spatial creep information of the slope is obtained by comparing the plurality of the first features, comparing the plurality of the second features, and comparing the plurality of the third features.

3. The landslide monitoring method according to claim 1, characterized in that: The first feature, the second feature and the third feature each include a plurality of first features, the plurality of first features are obtained based on optical remote sensing data corresponding to the first area of ​​the slope in a plurality of time units, the plurality of second features are obtained based on radar remote sensing data corresponding to the first area of ​​the slope in a plurality of time units, and the plurality of third features are obtained based on multi-source data corresponding to the first area of ​​the slope in a plurality of time units, and the first feature, the second feature and the third feature are respectively subjected to spatiotemporal analysis to obtain deformation information of the slope, further comprising: The time creep information of the slope is obtained by comparing the plurality of the first features, comparing the plurality of the second features, and comparing the plurality of the third features.

4. The landslide monitoring method according to claim 1, characterized in that: The landslide sensitivity factors include multiple factors, and the first model is obtained by the following steps: constructing a first matrix based on the plurality of landslide sensitivity factors, wherein the first matrix is ​​used to determine the weights of the landslide sensitivity factors; Based on the first matrix, obtaining a normalized weight of the landslide sensitivity factor; The first model LSI is obtained by the following formula: LSI=∑w i ×f i , Among them, f i represents the ith sensitivity factor, w i represents the normalized weight corresponding to the i-th landslide sensitivity factor, i is a natural number greater than 0, 0 <w i <1.

5. The landslide monitoring method according to claim 1, characterized in that: The obtaining of the first feature based on the optical remote sensing data comprises: The normalized difference vegetation index NDVI corresponding to the optical remote sensing data is obtained by the following formula: Among them, ρ NIR represents the reflectivity of the near-infrared band in the optical remote sensing data, ρ R represents the reflectivity of the red band in the optical remote sensing data; Performing partition processing on the optical remote sensing data to obtain a plurality of partition data; The first features corresponding to the plurality of partition data are obtained based on a pixel binary model, wherein the pixel binary model is determined by the following formula: Among them, NDVI soil Indicates the normalized difference vegetation index (NDVI) of pure vegetation corresponding to the partition data veg The normalized difference vegetation index of pure bare land corresponding to the partition data.

6. The landslide monitoring method according to claim 1, characterized in that: The second feature includes the surface deformation amount of the slope and the surface deformation rate of the slope. The second feature is obtained based on the radar remote sensing data, including: Extracting a time baseline and a space baseline from the radar remote sensing data based on the short baseline synthetic aperture radar interferometry (SBAS-InSAR) technology; Based on the time baseline and the space baseline, a connection graph of the radar remote sensing data is obtained; Performing interference measurement on the connection diagram to obtain a differential interference diagram; The differential interference pattern is subjected to phase unwrapping, singular value decomposition, and atmospheric phase removal operations to obtain the surface deformation amount and the surface deformation rate.

7. A landslide monitoring device, using the landslide monitoring method according to claim 1, characterized in that: include: An acquisition module is used to acquire multi-source data to be processed, wherein the multi-source data includes optical remote sensing data of the slope, radar remote sensing data of the slope, and digital elevation model DEM data of the slope; A feature extraction module, used to obtain a first feature based on the optical remote sensing data, a second feature based on the radar remote sensing data, and a third feature based on the multi-source data, wherein the first feature is used to represent the change in vegetation coverage of the slope, the second feature is used to represent the deformation information of the slope, and the third feature is used to represent the landslide sensitivity of the slope; The analysis module is used to perform spatiotemporal analysis on the first feature, the second feature and the third feature respectively to obtain creep information of the slope.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the landslide monitoring method according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the landslide monitoring method according to any one of claims 1 to 6 is implemented.

Images