Landslide remote sensing monitoring method, system and equipment based on time sequence

Abstract

The invention provides a landslide remote sensing monitoring method, system and equipment based on a time sequence, and the method comprises the steps: obtaining a long-time-sequence multispectral remote sensing data set of a target region, carrying out the preprocessing, and screening effective pixels; calculating normalized vegetation indexes NDVI of all effective pixels in the target area, and constructing a year-by-year NDVI data set of the target area; calculating an NDVI annual median value of each pixel, and constructing a time sequence NDVI curve of each pixel; carrying out mutation point detection on each pixel based on a linear regression method, calculating a residual error of an NDVI value, and carrying out comparative analysis to obtain a landslide monitoring result of each pixel; and correcting the landslide monitoring result to obtain a final landslide monitoring result in the target area. According to the method, the spatial and temporal distribution characteristics of landslides appearing in the target area can be quickly identified, the method has great significance in mastering landslide disaster monitoring and risk assessment in key areas, and the landslide disaster information extraction efficiency is remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the fields of remote sensing technology applications and natural disaster monitoring, and relates to a landslide remote sensing monitoring method, system, and equipment. In particular, it relates to a landslide remote sensing monitoring method, system, and equipment based on time series, providing new possibilities for realizing landslide monitoring using remote sensing technology. Background Art

[0002] In recent years, as one of the main geological disasters, the frequency of landslide disasters has increased significantly under the influence of global climate change. Landslide disasters are characterized by suddenness, concealment, and strong destructiveness. They often occur in complex terrain areas and usually result in serious consequences such as road interruptions, vehicle entrapment, and casualties. Therefore, continuously researching and adopting advanced technologies for landslide monitoring can timely detect and handle landslide hazards, contributing to ensuring the sustainable and stable development of the economy.

[0003] Traditional landslide monitoring technologies mainly rely on manual on-site surveys, which have disadvantages such as being easily affected by weather, limited measurement accuracy, high labor costs, and poor data real-time performance. Since landslide disasters often occur in remote areas with complex terrain, it is difficult for monitoring personnel to reach the site for surveys, with high potential risks, and it is also difficult to install and maintain monitoring equipment. Therefore, traditional landslide monitoring methods can no longer meet the monitoring requirements. In recent years, landslide monitoring technologies have gradually developed into a "sky-air-ground" collaborative three-dimensional monitoring mode, among which optical remote sensing monitoring technology is an important part of this collaborative three-dimensional monitoring mode. With the rapid development of optical satellite remote sensing technology, the generation speed of optical satellite remote sensing data has also increased rapidly, and using optical remote sensing images for landslide remote sensing monitoring shows great development potential.

[0004] Currently, using optical remote sensing data for landslide monitoring includes methods such as visual interpretation and machine recognition. Among them, the visual interpretation method requires relevant professional knowledge and experience, is easily affected by human factors, and has relatively low work efficiency; the machine recognition method is easily affected by various factors such as climate, terrain, and surface vegetation, and the accuracy of the recognition results cannot be guaranteed. Therefore, there is an urgent need to develop a remote sensing monitoring method for efficiently extracting landslide information using a large amount of remote sensing images. Summary of the Invention

[0005] The purpose of the present invention is to provide a landslide remote sensing monitoring method based on time series in view of the deficiencies of the prior art.

[0006] The present invention is achieved through the following technical solutions:

[0007] A landslide remote sensing monitoring method based on time series, comprising the following steps:

[0008] S1. Obtain the multi-spectral remote sensing data set of the target area for a long time series, perform preprocessing, and screen out valid pixels.

[0009] S3. Calculate the Normalized Difference Vegetation Index (NDVI) of all valid pixels in the target area, and construct the annual NDVI data set of the target area.

[0010] S3. Calculate the median NDVI within the year of each pixel according to the annual NDVI data set, and construct the time series NDVI curve of each pixel.

[0011] S4. Based on the NDVI curve, perform mutation point detection on each pixel using the linear regression method, calculate the residuals of the NDVI values and conduct comparative analysis to obtain the landslide monitoring results of each pixel, including information such as the landslide occurrence time and duration corresponding to the pixel.

[0012] S5. Conduct pseudo-landslide inspection on the pixels where landslides occur, and correct the landslide monitoring results according to the inspection results to obtain the final landslide monitoring results in the target area, including information such as the landslide occurrence location, occurrence range, occurrence time, and duration.

[0013] Further, step S1 is specifically as follows: Obtain the multi-spectral remote sensing images of the entire time series in the target area, perform geometric correction, radiometric correction, and atmospheric correction on the multi-spectral remote sensing images, and then use the cloud shadow monitoring results for pixel masking analysis to eliminate the pixels covered by clouds and shadows, and screen out the valid pixels of all images.

[0014] Further, step S3 is specifically as follows: Use the annual NDVI data set to calculate the median NDVI within the year raster data of each pixel, construct the annual seamless NDVI monitoring raster layer according to the median NDVI within the year raster data, and extract the median NDVI of the time series of each pixel from the seamless NDVI monitoring raster layer, so as to construct the NDVI curve of the time series of each pixel.

[0015] Further, in step S3, the moving window method is used to perform mean smoothing processing on the NDVI curve.

[0016] Further, step S4 is specifically as follows:

[0017] Based on the NDVI curve, perform mutation point detection on each pixel using the linear regression method to obtain several mutation points;

[0018] Use the mutation points as segmentation points to segment the time series between two adjacent mutation points as sub-time series;

[0019] Calculate the NDVI residuals for two adjacent years of each mutation point to obtain the residual results of each sub-time series. Analyze the residual results to obtain the landslide occurrence time corresponding to each pixel.

[0020] Further, the calculation method of the NDVI residual is as follows:

[0021]

[0022] Among them, is the NDVI residual, is the median NDVI value in the year of the mutation point, is the median NDVI value in the year following the year of the mutation point.

[0023] Further, the analysis of the residual results to obtain the landslide occurrence time corresponding to each pixel is specifically as follows: The year corresponding to the maximum value of the NDVI residual is the year when the landslide starts, the following year is the year when the landslide occurs, and the year corresponding to the minimum value of the NDVI residual is the year when the landslide is completed.

[0024] Further, step S5 is specifically as follows: Calculate the difference between the median NDVI values of the year when the landslide starts and the year when the landslide is completed for each pixel. If the difference is less than the preset threshold, it is determined as a false landslide, and the landslide monitoring results are corrected according to the false landslide inspection results to obtain the final landslide monitoring results in the target area.

[0025] A time-series-based remote sensing landslide monitoring system includes:

[0026] Data processing module: Used to obtain the multi-spectral remote sensing data set of the entire time series of the target area, perform preprocessing, and screen valid pixels;

[0027] NDVI calculation module: Used to calculate the normalized vegetation index NDVI of all valid pixels in the target area and construct the annual NDVI data set of the target area;

[0028] Curve construction module: Used to calculate the median NDVI value within the year of each pixel according to the annual NDVI data set and construct the time-series NDVI curve of each pixel;

[0029] Landslide detection module: Used to detect mutation points for each pixel based on the linear regression method according to the NDVI curve, calculate the residuals of the NDVI values and perform comparative analysis to obtain the landslide monitoring results of each pixel;

[0030] Result correction module: Perform false landslide inspection on the pixels where landslides occur, and correct the landslide monitoring results according to the inspection results to obtain the final landslide monitoring results in the target area.

[0031] A computer device includes:

[0032] One or more processors;

[0033] A memory for storing one or more programs;

[0034] When the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned landslide remote sensing monitoring method based on time series.

[0035] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0036] For traditional landslide detection and identification, on-site manual surveys are generally required, which are not only time-consuming and laborious, but also the safety cannot be guaranteed. For areas with complex terrain or areas that cannot be reached due to other reasons, traditional detection methods are even more powerless. In view of the target area without human activities and the situation where on-site surveys cannot be carried out, the present invention proposes a method for extracting landslide information using long-time series multispectral images. This method has the advantages of not requiring on-site surveys, reducing the influence of weather, being able to utilize multi-temporal data, and not requiring manual screening of images. It can quickly identify the spatio-temporal distribution characteristics of landslides that have occurred in the target area, which is of great significance for mastering the monitoring and risk assessment of landslide disasters in key areas and significantly improving the efficiency of extracting landslide disaster information. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 is the flow chart of the landslide remote sensing monitoring method based on time series in the embodiment of the present invention.

[0038] Figure 2 is the statistical schematic diagram of the effective observation values of each pixel in the study area in a typical year (2022) in the embodiment of the present invention.

[0039] Figure 3 is the NDVI raster layer of a typical year (2022) obtained by calculating the median value of each pixel within the year in the study area in the embodiment of the present invention.

[0040] Figure 4 is the NDVI curve of the time series of a typical pixel in the embodiment of the present invention.

[0041] Figure 5 is the sub-time series obtained by detecting the mutation points and segmenting the NDVI curve of the time series of a typical pixel in the embodiment of the present invention.

[0042] Figure 6 is an example of the monitoring result of time series landslide remote sensing in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] The technical solution of the present invention will be further described in detail below with reference to the drawings and specific examples.

[0044] A landslide remote sensing monitoring method based on time series, comprising the following steps:

[0045] S1. Obtain a multi-spectral remote sensing data set of a target area with a long time series, perform geometric correction, radiometric correction, and atmospheric correction on the multi-spectral remote sensing image, and then use the cloud shadow monitoring results to perform pixel masking analysis to eliminate cloud-covered and shadow-covered pixels, and screen out the effective pixels of all images;

[0046] S2. Calculate the Normalized Difference Vegetation Index (NDVI) of all effective pixels in the target area, and construct an annual NDVI data set for the target area; the NDVI calculation method is as follows:

[0047]

[0048] Among them, NIR and Red are the surface reflectance of the near-infrared band and the red band respectively, and n is the number represented by the image. The NDVI value is between -1 and 1. When it is negative, it indicates no vegetation coverage, and when it is positive, it indicates vegetation coverage. The larger the value, the better the vegetation growth. Assign the NDVI (negative value) of the non-vegetation-covered area to 0. Therefore, the value of the vegetation index NDVI ranges from 0 to 1.

[0049] S3. Use the annual NDVI data set to calculate the median NDVI raster data within the year for each pixel. The median NDVI raster data within the year The calculation method is as follows:

[0050]

[0051] Construct an annual seamless NDVI monitoring raster layer based on the median NDVI raster data within the year, extract the median NDVI of the time series of each pixel from the seamless NDVI monitoring raster layer, and thus construct the NDVI curve of the time series of each pixel. Further, use the moving window method (the window size is set to 3 years) to perform mean smoothing processing on the annual time series curve to make the trend of the curve smoother and more reasonable.

[0052] S4. Based on the NDVI curve, perform mutation point detection on each pixel using the linear regression method to obtain several mutation points; use the mutation points as segmentation points to segment the time series between two adjacent mutation points as a sub-time series; calculate the NDVI residuals of two adjacent years of each mutation point to obtain the residual results of each sub-time series, and analyze the residual results to obtain the landslide occurrence time corresponding to each pixel.

[0053] The calculation method of the NDVI residual is:

[0054]

[0055] Among them, is the NDVI residual, is the median value of NDVI in the year when the mutation point occurs, is the median value of NDVI in the year next to the year when the mutation point occurs.

[0056] The year corresponding to the maximum value of the NDVI residual is the starting year of the landslide, and the year corresponding to the minimum value of the NDVI residual is the completion year of the landslide.

[0057] S5. Conduct a pseudo-landslide check on the pixels where landslides occur, and correct the landslide monitoring results according to the check results to obtain the final landslide monitoring results in the target area. Specifically: calculate the difference between the median value of NDVI in the starting year and the completion year of the landslide for each pixel. If the difference is less than the preset threshold, it is determined as a pseudo-landslide. Correct the landslide monitoring results according to the pseudo-landslide check results, and integrate the landslide analysis results of each pixel to obtain the final landslide monitoring results in the target area, including information such as the location, scope, occurrence time, and duration of the landslide.

[0058] Through the above work process, the time-series remote sensing monitoring of landslides can be completed, and the landslide results including information such as the location, scope, occurrence time, and duration of the landslides in the study area can be obtained.

[0059] In a specific embodiment of the present invention, taking a certain place as the study area, as Figure 1-6 shown, specifically includes the following steps:

[0060] (1)Obtain long-term multi-spectral remote sensing images of a certain area. In this example, Landsat multi-spectral data is used as the data source. First, obtain the time-series Landsat multi-spectral images covering the study area. In this example, a total of 287 Landsat images covering the study area from 1985 to 2022 are obtained. Further, preprocessing operations such as geometric correction, radiometric correction, and atmospheric correction are carried out on the obtained images. Using the cloud shadow monitoring results, the pixels containing cloud shadows are masked (for the specific method, refer to Steve Foga, Pat L. Scaramuzza, Song Guo, Zhe Zhu, Ronald D. Dilley, Tim Beckmann, Gail L. Schmidt, John L. Dwyer, M. Joseph Hughes, Brady Laue, Cloud detection algorithm comparison and validation for operational Landsat data products, Remote Sensing of Environment, Volume 194, 2017, Pages 379-390, ISSN 0034-4257), and the effective pixels of the surface reflectance of all remote sensing images are screened out. After statistical analysis of the effective observation numbers of each pixel, it is found that the study area of this embodiment is a cloudy and rainy area. During the period from 1985 to 2022, in the remotely sensed images collected, the proportion of effective observation values of each pixel is greatly affected by the weather. Among them, in 2022, the maximum number of effective observation values of each pixel in the study area is 12, and the minimum is 5 ( Figure 2 ).

[0061] (2)Calculate the NDVI for each effective pixel of each image, and construct an annual NDVI dataset for the study area with the natural year as the unit. The specific calculation formula is as follows:

[0062]

[0063] Among them, NIR and Red are the surface reflectances of the near-infrared band and the red band respectively, and n is the image data within the year. The NDVI value is between -1 and 1. When it is negative, it indicates no vegetation cover, and when it is positive, it indicates vegetation cover. The larger the value, the better the vegetation growth. The NDVI (negative value) of the vegetation-free area is assigned 0. Therefore, the value of the vegetation index NDVI ranges from 0 to 1.

[0064] (3)Construct a seamless monitoring raster layer of annual NDVI for the study area. The specific method is as follows: Using the annual NDVI dataset, calculate the median value of NDVI within each pixel in the study area for each year to construct the seamless monitoring raster layer of annual NDVI for the study area ( Figure 3 ), and the method is as follows:

[0065]

[0066] Furthermore, obtain the time series NDVI curve for each pixel. First, use the above-mentioned seamless monitoring raster layer of annual NDVI to construct the time series NDVI curve for each pixel in the study area, Figure 4 which is the time series NDVI curve of a typical pixel selected in this embodiment. Further, adopt the moving window method (the window size is set to 3 years) to perform mean smoothing on the annual time series curve to make the trend of the curve smoother and more reasonable.

[0067] Furthermore, conduct landslide detection for each pixel. The specific method is as follows: First, adopt the linear regression method to detect the mutation points for each pixel in the study area, and segment the time series between two adjacent mutation points as a sub-time series to obtain the sub-time series. Figure 5 is the time series NDVI curve of a typical pixel after mutation point detection, and 6 mutation points are obtained. These mutation points divide the time series NDVI curve into 7 sub-time series ( Figure 5 ).

[0068] Furthermore, obtain the preliminary landslide monitoring results for the study area. The specific method is as follows: Calculate the NDVI residuals for the two years before and after the occurrence of the mutation points to obtain the residual results for each sub-time series. The method is as follows:

[0069]

[0070] where, and represent the NDVI values for the year of the change point and the following year respectively.

[0071] In this embodiment, the residual values for the two years before and after the 6 mutation points are calculated as follows:

[0072] Change point number Corresponding year Change point NDVI value Next year's NDVI value Residual P1 1990 0.746 0.588 0.158 P2 1991 0.588 0.607 -0.019 P3 1994 0.771 0.748 0.023 P4 2000 0.720 0.577 0.143 P5 2003 0.261 0.272 -0.011

[0073] Furthermore, analyze the residual results to obtain the landslide occurrence time and completion time corresponding to each pixel, and obtain the preliminary landslide monitoring results of the study area. The specific method is as follows: The year corresponding to the maximum residual value is the year when the landslide starts (pstart). If the next year (pnext) of this pixel is a valid pixel, then pnext is determined as the year when the landslide occurs. The year corresponding to the minimum residual value (pend) is the year when the landslide is completed. Preliminary analysis of the typical pixel in this embodiment reveals that the NDVI mutation of this pixel is divided into three stages ( Figure 5 ): In the first stage (S2), the year corresponding to the maximum residual value is P1, and the year corresponding to the minimum residual value is P2. It can be preliminarily determined that a landslide occurred in this stage. The landslide occurrence time is the next year of P1, and the end time is P2. In the second stage (S3), the year corresponding to the maximum residual value is P3, and the year corresponding to the minimum residual value is P2. It can be seen that the NDVI value continuously increases in this stage, and it is impossible for a landslide to occur. In the third stage (S5), the year corresponding to the maximum residual value is P4, and the year corresponding to the minimum residual value is P5. It can be preliminarily determined that a landslide occurred in this stage. The landslide occurrence time is the next year of P4, and the end time is P5.

[0074] Furthermore, conduct pseudo-landslide inspection to correct the preliminary landslide monitoring results. The specific method is as follows: First, conduct pseudo-landslide inspection. The specific method is: Calculate the change degree of the NDVI median value from the landslide start year (pstart) to the landslide completion year (pend) of each pixel. If the difference is less than 0.2, then this pixel is determined as a pseudo-landslide. In the preliminary analysis results of the typical pixel in this embodiment, the NDVI difference between the landslide start year (P1) and the landslide completion year (P2) in the S2 stage is 0.158, and the landslide in this stage is a pseudo-landslide. The NDVI difference between the landslide start year (P4) and the landslide completion year (P5) in the S5 stage is 0.459, and this stage is determined as a landslide. The landslide occurrence time is the next year of P4 (2001), and the end time is P5 (2003).

[0075] Furthermore, obtain the spatio-temporal distribution characteristic information of the landslides in the study area. The specific method is to use the pseudo-landslide inspection results to eliminate the pseudo-landslide information, correct the preliminary landslide monitoring results of all pixels, and obtain the landslide monitoring results of the study area including information such as the landslide occurrence location, occurrence scope, occurrence time, and duration ( Figure 6 ).

[0076] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0077] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices produce means for implementing the functions specified in Figure 1 one or more of the flows Figure 1 or blocks.

[0078] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in Figure 1 one or more of the flows Figure 1 or blocks.

[0079] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one or more of the flows Figure 1 or blocks.

[0080] The above are only the preferred embodiments of the present invention. Although the present invention has been disclosed above with preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make many possible changes and modifications to the technical solution of the present invention, or modify it into equivalent embodiments with equivalent changes, without departing from the scope of the technical solution of the present invention. Therefore, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A landslide remote sensing monitoring method based on time series, characterized in that, It includes the following steps: S1. Obtain the multi-spectral remote sensing data set of the target area with a long time series, perform preprocessing, and screen out valid pixels; S2. Calculate the Normalized Difference Vegetation Index (NDVI) of all valid pixels in the target area, and construct the annual NDVI data set of the target area; S3. Calculate the median NDVI value within each year for each pixel based on the annual NDVI data set, and construct the time series NDVI curve for each pixel; S4. Based on the NDVI curve, perform mutation point detection on each pixel using the linear regression method, calculate the residuals of the NDVI values and conduct comparative analysis to obtain the landslide monitoring results for each pixel; S5. Conduct pseudo-landslide inspection on the pixels where landslides occur, and correct the landslide monitoring results according to the inspection results to obtain the final landslide monitoring results in the target area.

2. The landslide remote sensing monitoring method based on time series according to claim 1, characterized in that Specifically, step S1 is: Obtain the multi-spectral remote sensing images of the entire time series in the target area, perform geometric correction, radiometric correction and atmospheric correction on the multi-spectral remote sensing images, and then use the cloud shadow monitoring results for pixel masking analysis to screen out the valid pixels of all images.

3. The landslide remote sensing monitoring method based on time series according to claim 1, characterized in that Specifically, step S3 is: Use the annual NDVI data set to calculate the median NDVI raster data within each year for each pixel, construct the annual seamless NDVI monitoring raster layer according to the median NDVI raster data within each year, and extract the median NDVI of the time series of each pixel from the seamless NDVI monitoring raster layer, so as to construct the NDVI curve of the time series of each pixel.

4. The landslide remote sensing monitoring method based on time series according to claim 3, characterized in that In step S3, the moving window method is used to perform mean smoothing processing on the NDVI curve.

5. The landslide remote sensing monitoring method based on time series according to claim 1, characterized in that Specifically, step S4 is: Based on the NDVI curve, perform mutation point detection on each pixel using the linear regression method to obtain several mutation points; Use the mutation points as segmentation points to segment the time series between two adjacent mutation points as sub-time series; Calculate the NDVI residuals of two adjacent years for each mutation point to obtain the residual results of each sub-time series, analyze the residual results, and obtain the landslide occurrence time corresponding to each pixel.

6. The landslide remote sensing monitoring method based on time series according to claim 5, characterized in that The calculation method of the NDVI residual is: ， Among them, is the NDVI residual, is the median NDVI value in the year when the mutation point occurs, is the median NDVI value in the year next to the year when the mutation point occurs.

7. The landslide remote sensing monitoring method based on time series according to claim 5, characterized in that The analysis of the residual results to obtain the landslide occurrence time corresponding to each pixel is specifically: The year corresponding to the maximum value of the NDVI residual is the year when the landslide starts, the next year is the year when the landslide occurs, and the year corresponding to the minimum value of the NDVI residual is the year when the landslide is completed.

8. The landslide remote sensing monitoring method based on time series according to claim 1, wherein Specifically, step S5 is: Calculate the difference between the median NDVI values of the year when the landslide starts and the year when the landslide is completed for each pixel. If the difference is less than the preset threshold, it is determined as a pseudo-landslide, and the landslide monitoring results are corrected according to the pseudo-landslide inspection results to obtain the final landslide monitoring results in the target area.

9. A landslide remote sensing monitoring system based on time series, characterized in that, It includes: Data processing module: Used to obtain the multi-spectral remote sensing data set of the entire time series in the target area, perform preprocessing, and screen out valid pixels; NDVI calculation module: Used to calculate the Normalized Difference Vegetation Index (NDVI) of all valid pixels in the target area, and construct the annual NDVI data set of the target area; Curve construction module: used to calculate the median NDVI within the year for each pixel based on the annual NDVI dataset, and construct the time-series NDVI curve for each pixel; Landslide detection module: used to detect mutation points for each pixel based on the linear regression method according to the NDVI curve, calculate the residuals of the NDVI values and conduct comparative analysis to obtain the landslide monitoring results for each pixel; Result correction module: conduct pseudo-landslide inspection on the pixels where landslides occur, and correct the landslide monitoring results according to the inspection results to obtain the final landslide monitoring results within the target area.

10. A computer device, characterized in that, Comprising: One or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the time-series-based remote sensing landslide monitoring method according to any one of claims 1-8.

Images