Forest landslide and collapse area detection method based on unmanned aerial vehicle visual remote sensing

Abstract

A forest landslide collapse area detection method based on unmanned aerial vehicle visual remote sensing, characterized in that it comprises the following steps: 1. Constructing a forest landslide collapse area detection system based on unmanned aerial vehicle visual remote sensing; 2. Unmanned aerial vehicle shooting remote sensing image data; 3. The image preprocessing module performs preprocessing operation on the historical satellite image data and the remote sensing image data, and obtains the remote sensing image and the historical image; 4. The image alignment module matches the feature points of the remote sensing image and the historical image, calculates the transformation matrix, and completes the alignment of the images before and after the landslide through the transformation matrix; 5. The landslide area recognition module recognizes the landslide area; 6. The landslide area recognition module extracts the landslide area contour; 7. The landslide area calculation module calculates the average slope of the landslide area, and combines the planar projection area and the average slope to calculate the actual area of the landslide inclined surface. Effect: It can quickly and accurately realize the detection of the forest landslide collapse area.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of forest landslide detection, and particularly relates to a forest landslide collapse area detection method based on unmanned aerial vehicle vision remote sensing. BACKGROUND

[0002] Forest landslide is a common geological disaster that poses a serious threat to forest ecosystems and human society. Forest landslide not only leads to large-scale destruction of vegetation cover, but also can trigger secondary disasters such as mudslides, posing a significant risk to the surrounding environment and residents' safety. Timely and accurate detection and monitoring of the occurrence range and area of forest landslide is of great significance for assessing disaster impact, developing prevention and control measures, and guiding ecological restoration work.

[0003] Accurate measurement of landslide area is a key link in disaster assessment. It not only reflects the scale and intensity of the landslide directly, but also provides important basis for subsequent ecological restoration and geological control. Accurate landslide area data can help decision-makers better allocate resources, develop targeted restoration strategies, and evaluate restoration effectiveness. In addition, time series analysis of landslide area data can reveal the development trend of landslides, providing scientific support for the establishment and improvement of early warning systems.

[0004] Traditional methods for detecting the area of forest landslide mainly include field measurement and satellite remote sensing. Field measurement method requires professional personnel to enter the landslide area for field investigation and measurement. Although this method can obtain relatively accurate data, it has the following shortcomings: first, the operation process is time-consuming and laborious, especially in complex terrain and inaccessible forest areas; second, field measurement may face safety risks, especially when the landslide has just occurred or is still ongoing; finally, field measurement is inefficient and difficult to meet the rapid assessment needs of large-scale or multiple landslide disasters.

[0005] Although satellite remote sensing technology overcomes some limitations of field measurement, it also has obvious shortcomings. The acquisition of high-resolution satellite images is costly and is limited by orbital periods and weather conditions, making it difficult to achieve timely monitoring of landslide areas. Low-resolution satellite images, while easier to obtain, often lack sufficient spatial resolution to accurately identify and measure small landslides. In addition, satellite remote sensing is limited in application in dense forest areas, making it difficult to penetrate dense canopies to obtain ground information.

[0006] Therefore, the existing technology has the following main defects in the detection of forest landslide collapse area: the field measurement method is time-consuming, laborious, has safety risks, and is inefficient; the satellite remote sensing method is limited by high cost, low timeliness, and application limitations in complex terrain and dense forest areas. These problems seriously restrict the rapid assessment and effective management of forest landslide disasters. SUMMARY

[0007] The application provides a forest landslide collapse area detection method based on unmanned aerial vehicle visual remote sensing, which can quickly, accurately, safely and economically and efficiently realize accurate detection of the forest landslide collapse area.

[0008] To achieve the above object, the application provides a forest landslide collapse area detection method based on unmanned aerial vehicle visual remote sensing, which is characterized by comprising the following steps.

[0009] Step 1: system construction: a forest landslide collapse area detection system based on unmanned aerial vehicle visual remote sensing is constructed, the forest landslide collapse area detection system comprises an unmanned aerial vehicle, a multispectral camera is carried on the unmanned aerial vehicle, and the multispectral camera is sequentially connected with an image preprocessing module, an image alignment module, a landslide area identification module and a landslide area calculation module.

[0010] Step 2: data acquisition: the unmanned aerial vehicle carries the multispectral camera to shoot remote sensing images of the forest landslide area, obtains remote sensing image data, and transmits the remote sensing image data to the image preprocessing module.

[0011] Step 3: image preprocessing: the image preprocessing module obtains historical satellite image data of the forest landslide area, and performs a preprocessing operation on the historical satellite image data and the remote sensing image data, to obtain standard remote sensing images and historical images, and transmits the standard remote sensing images and the historical images to the image alignment module.

[0012] Step 4: image alignment: the image alignment module extracts feature points in the remote sensing images and the historical images, matches the feature points, then calculates a transformation matrix according to the matched feature points, and maps the remote sensing images and the historical images to the same spatial reference system through the transformation matrix, to complete alignment of images before and after the landslide.

[0013] Step 5: landslide area identification: the landslide area identification module identifies a landslide area by calculating a ratio vegetation index RVI of the remote sensing images and the historical images.

[0014] Step 6: image post-processing: the landslide area identification module removes small noises and connects dispersed landslide areas by using erosion and expansion operations of OpenCV, and extracts a landslide area contour.

[0015] Step 7: landslide area calculation: the landslide area calculation module generates a digital elevation model DEM by using a GDAL tool to perform coordinate conversion, calculates an average slope of the landslide area, and then combines a planar projection area and the average slope to calculate an actual area of a landslide inclined surface, to obtain the forest landslide collapse area.

[0016] Through the above design, first, the unmanned aerial vehicle is used to carry a multispectral camera to obtain remote sensing images of the landslide area, and historical satellite images before the landslide are collected. Then, data preprocessing is performed, including radiation calibration and image correction operations. Next, the SIFT algorithm is used to align the images before and after the landslide to ensure that the spatial positions of the images are consistent. Subsequently, the RVI (Ratio Vegetation Index) is used to identify the bare soil area after the landslide, and the landslide area is determined according to the change in the RVI value. Finally, the erosion and expansion operations are performed on the landslide area to extract the final landslide area contour, realizing the accurate identification and area calculation of the landslide area. Through this method, the forest landslide collapse area can be efficiently, accurately, safely and economically identified and calculated, providing a scientific basis for disaster assessment and the development of prevention measures.

[0017] As a preferred: in the step 2, the unmanned aerial vehicle is used to carry a multispectral camera to take multiple overlapping remote sensing images of the forest landslide area, with a heading overlap rate and a lateral overlap rate of not less than 80%, for subsequent reconstruction of the three-dimensional point cloud of the landslide area to obtain the elevation model.

[0018] At the same time, historical satellite remote sensing images before the landslide are collected to provide basic data for subsequent landslide area reconstruction and comparative analysis.

[0019] As a preferred: in the step 3, the image preprocessing module performs preprocessing operations on the remote sensing image data and historical satellite image data, and the specific steps are as follows:

[0020] Step 31: radiation calibration: the image preprocessing module converts the remote sensing image data and historical satellite image data stored in the form of digital numerical value DN into the form of radiation luminance value through a conversion formula, and the conversion formula is as follows:

[0021] L = (DN - DN offset ) × Gain

[0022] Wherein, L is the radiation luminance value, DN is the original digital numerical value, DN offset is the offset value of the sensor, and Gain is the gain coefficient;

[0023] The original image data captured by the remote sensor is stored in the form of digital numerical value DN, which is converted into radiation luminance value through a conversion formula, i.e. the radiation energy per unit area, unit solid angle and unit wavelength range, which is convenient for subsequent image correction steps to correct the atmospheric apparent reflectance to the ground reflectance.

[0024] Step 32: image correction: the image preprocessing module performs processing operations on the remote sensing image data and historical satellite image data in the form of radiation luminance value, including but not limited to geometric correction, image fusion, image cropping and shadow removal, and then uses the FLAASH atmospheric correction model to correct the radiation reflectance to the ground reflectance.

[0025] In the process of remote sensing imaging, the instability of the sensor and the water vapor, aerosol scattering effect and other factors may cause the geometric distortion and radiation distortion of the image. These factors have a serious impact on the quality of remote sensing image. For the geometric distortion problem caused by the instability of the sensor, geometric correction (including coarse correction, geometric fine correction and orthographic correction), image fusion, image cropping and shadow removal and other operations are used to process and improve the quality of remote sensing image; for the radiation distortion caused by water vapor, aerosol scattering effect and other factors, the FLAASH atmospheric correction model is used to remove the atmospheric influence, and the atmospheric apparent reflectance is calculated, and then the atmospheric apparent reflectance is corrected to the ground reflectance, which reflects the true reflectance of the ground.

[0026] As preferred: the geometric correction includes coarse correction, geometric fine correction and orthographic correction.

[0027] As preferred: in the step 4, the image alignment module uses the SIFT scale invariant feature transform method to detect the feature points in the remote sensing image and the historical image, and then matches the feature points of the remote sensing image and the historical image through the nearest neighbor search algorithm to obtain a set of feature point pairs.

[0028] The purpose of remote sensing image alignment is to accurately align the satellite image before the landslide with the unmanned aerial vehicle remote sensing image after the landslide, so as to ensure that the same geographical position has consistent spatial position in the images at different times. The SIFT scale invariant feature transform method is used to detect the feature points in the historical image and the remote sensing image, and the feature descriptor of each feature point is generated, and the nearest neighbor search algorithm is used to find the corresponding feature point pairs in the historical image before the landslide and the remote sensing image after the landslide. The transformation matrix is calculated through the matched feature point pairs, which maps the images before and after the landslide to the same spatial reference system.

[0029] As preferred: in the step 5, the landslide area identification module reads the remote sensing image using the OpenCV, a cross-platform computer vision and machine learning software library, and extracts the data of the red light band and the near-infrared band; the calculation expression of the ratio vegetation index RVI of the remote sensing image is as follows:

[0030]

[0031] Where, I RED is the red light band data value of the corresponding pixel in the remote sensing image, I NIR is the near-infrared band data value of the corresponding pixel in the remote sensing image.

[0032] The ratio vegetation index RVI is calculated by the ratio of red band and near-infrared band, which can be used to distinguish vegetation and soil, and then the ratio vegetation index RVI is used to identify the debris area generated after the landslide.

[0033] An RVI threshold T is set to distinguish soil and non-soil areas, and the RVI value in the range of 0-2 is identified as bare soil and rock area, and the RVI value in the range of 2-8 is identified as green vegetation.

[0034] The occurrence of forest landslide will cover the forest vegetation area with bare soil, and the area that is vegetation before the landslide will have a decrease in RVI value after the landslide. For the case that the forest vegetation area is covered by landslide debris, the RVI value before the landslide will be between 2 and 8, and the RVI value after the landslide will be between 0 and 2. Therefore, the area of forest landslide occurrence is determined according to the change of RVI value.

[0035] As preferred: in the step 6, the landslide area identification module uses the erosion function in OpenCV to perform erosion and dilation processing on the identified landslide area, and the specific steps are as follows:

[0036] Step 61: select a 3X3 erosion convolution kernel, use the cv2.erode() function in OpenCV to perform erosion operation on the landslide area image, and get the eroded image;

[0037] The erosion operation can remove small noise points in the landslide area and eliminate small patches that are not connected to the main landslide area. The erosion operation replaces the pixel values in the region with the local minimum value by sliding the convolution kernel in the image, thereby reducing the region.

[0038] Step 62: select a 3X3 dilation convolution kernel, use the cv2.dilate() function in OpenCV to perform dilation operation on the eroded image, and get the dilated image;

[0039] Dilation can expand and connect dispersed landslide areas to make them more coherent, while restoring the main landslide area that may have been excessively reduced during the erosion process. The dilation operation replaces the pixel values in the region with the local maximum value by sliding the convolution kernel in the image, thereby expanding the region.

[0040] Step 63: use the contour extraction function cv2.findContours() in OpenCV to extract the contour of the dilated image, get the landslide area contour, and facilitate subsequent calculation of the area of the landslide area or other analysis.

[0041] As preferred: in the step 7, the landslide area calculation module calculates the forest landslide collapse area, and the specific steps are as follows:

[0042] Step 71: The landslide area calculation module uses the `gdaltransform` tool in the open source raster spatial data conversion library GDAL to convert the pixel coordinates (row, column) in the remote sensing image to geographic coordinates (longitude, latitude); uses the `ogr2ogr` tool in the open source raster spatial data conversion library GDAL to save the converted landslide boundary geographic coordinates in vector format, such as Shapefile or GeoJSON, for further analysis and visualization in GIS software.

[0043] Step 72: Use the three-dimensional reconstruction open source software library OpenMVG to estimate camera pose and generate sparse point cloud, after generating sparse point cloud, further optimize the reconstruction model and generate dense point cloud, then use GDAL tool to convert the dense point cloud into digital elevation model DEM; Each point of the dense point cloud contains X, Y and Z coordinates, use `gdal_grid` command in GDAL to interpolate each point of the dense point cloud to generate regular DEM grid data;

[0044] Step 73: According to the DEM grid data, use the gdaldem command in the GDAL tool to calculate the slope; By creating a mask, extract the slope data of the landslide area in the remote sensing image, after extracting all the slope values in the landslide area, calculate the average value of all the slope values, the slope average value calculation formula is as follows:

[0045]

[0046] Where, N is the total number of pixels in the landslide area, θ i is the slope value of the i-th pixel, unit: degree;

[0047] Step 74: According to the landslide area contour, get the coordinates (x i ,y i ) on the landslide area contour, then calculate the projection area of the landslide area on the horizontal plane, the calculation formula is as follows:

[0048]

[0049] Where, (x i ,y i ) is the plane coordinates of the i-th vertex of the polygon, n is the number of polygon vertices;

[0050] Then calculate the actual area of the landslide slope, that is, the forest landslide collapse area, using the average slope, the calculation formula is as follows:

[0051]

[0052] Where, A 平面θ is the calculated planar projected area, and θ is the average slope of the landslide area.

[0053] The beneficial effects of this invention are: it can efficiently, accurately, safely and cost-effectively identify and calculate the area of ​​forest landslides and collapses, and achieve precise monitoring of the extent of forest landslides and collapses, aiming to provide a scientific basis for landslide disaster assessment, prevention and control measures formulation and ecological restoration planning. Attached Figure Description

[0054] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation

[0055] The present invention will be further described in detail below with reference to the accompanying drawings and specific examples. The following embodiments or drawings are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0056] like Figure 1 As shown: A method for detecting the area of ​​forest landslide collapse based on UAV visual remote sensing, including the following steps:

[0057] Step 1: System Construction: Construct a forest landslide collapse area detection system based on UAV visual remote sensing. The forest landslide collapse area detection system includes a UAV, which is equipped with a multispectral camera. The multispectral camera is sequentially connected to an image preprocessing module, an image alignment module, a landslide area identification module, and a landslide area calculation module.

[0058] Step 2: Data Acquisition: The UAV is equipped with a multispectral camera to capture remote sensing images of the forest landslide area, obtain remote sensing image data, and transmit it to the image preprocessing module;

[0059] Step 3: Image preprocessing: The image preprocessing module acquires historical satellite image data of the forest landslide area, and performs preprocessing operations on the historical satellite image data and remote sensing image data to obtain standard remote sensing images and historical images, which are then passed to the image alignment module.

[0060] Step 4: Image Alignment: The image alignment module extracts feature points from the remote sensing image and the historical image, matches the feature points of the two, calculates the transformation matrix based on the matched feature point pairs, and maps the remote sensing image and the historical image to the same spatial reference system through the transformation matrix to complete the alignment of the images before and after the landslide.

[0061] Step 5: Landslide Area Identification: The landslide area identification module identifies landslide areas by calculating the ratio of the remote sensing image to the historical image, the vegetation index (RVI).

[0062] Step 6: Image post-processing: the landslide area recognition module uses the erosion and dilation operations of OpenCV to remove small noise and connect scattered landslide areas, and extracts the landslide area contour.

[0063] Step 7: Landslide area calculation: the landslide area calculation module uses the GDAL tool to perform coordinate conversion, generates a digital elevation model DEM, calculates the average slope of the landslide area, and then combines the planar projection area and the average slope to calculate the actual area of the landslide inclined surface, obtaining the forest landslide collapse area.

[0064] In step 2, the unmanned aerial vehicle carries a multispectral camera to take multiple overlapping remote sensing images of the forest landslide area, with a heading overlap rate and a lateral overlap rate of not less than 80%. The three-dimensional point cloud of the landslide area is reconstructed to obtain the elevation model; in addition, historical satellite image data before the forest landslide occurs is also obtained.

[0065] In step 3, the image preprocessing module performs preprocessing operations on the remote sensing image data and historical satellite image data, and the specific steps are as follows:

[0066] Step 31: Radiometric calibration: the image preprocessing module converts the remote sensing image data and historical satellite image data stored in digital numerical value DN form into radiometric value form through a conversion formula, and the conversion formula is as follows:

[0067] L = (DN - DN offset ) x Gain

[0068] Where L is the radiometric value, DN is the original digital value, DN offset is the offset value of the sensor, and Gain is the gain coefficient;

[0069] Step 32: Image correction: the image preprocessing module performs geometric correction, image fusion, image cropping, shadow removal, and other processing operations on the remote sensing image data and historical satellite image data in radiometric value form, and then uses the FLAASH atmospheric correction model to correct the radiometric reflectance to the ground reflectance.

[0070] The geometric correction includes coarse correction, geometric fine correction, and orthographic correction.

[0071] In step 4, the image alignment module uses the SIFT scale-invariant feature transform method to detect feature points in the remote sensing image and historical image, and then matches the feature points in the remote sensing image and historical image through the nearest neighbor search algorithm to obtain a set of feature point pairs.

[0072] In the step 5, the landslide area identification module reads the remote sensing image using the cross-platform computer vision and machine learning software library OpenCV, extracts the data of the red light band and the near-infrared band; and the calculation expression of the ratio vegetation index RVI of the remote sensing image is as follows:

[0073]

[0074] wherein, I RED is the red light band data value of the corresponding pixel in the remote sensing image, I NIR is the near-infrared band data value of the corresponding pixel in the remote sensing image.

[0075] In the embodiment, the RVI threshold T is set to 2, which is used to distinguish soil and non-soil areas, the RVI value in the range of 0-2 is identified as bare soil and rock area, and the RVI value in the range of 2-8 is identified as green vegetation.

[0076] The forest landslide will cover the forest vegetation area with bare soil, and the area with RVI value representing vegetation before the landslide will have a decreased RVI value after the landslide. For the case that the forest vegetation area is covered by landslide debris, the RVI value before the landslide will be between 2 and 8, and the RVI value after the landslide will be between 0 and 2. Therefore, the area where the forest landslide occurs is determined according to the change of the RVI value.

[0077] In the step 6, the erosion and expansion processing of the identified landslide area is performed by the erosion function in OpenCV, and the specific steps are as follows:

[0078] Step 61: a 3X3 erosion convolution kernel is selected, and the cv2.erode() function in OpenCV is used to perform erosion operation on the landslide area image to obtain the eroded image;

[0079] Step 62: a 3X3 expansion convolution kernel is selected, and the cv2.dilate() function in OpenCV is used to perform expansion operation on the eroded image to obtain the expanded image;

[0080] Step 63: the contour extraction function cv2.findContours() in OpenCV is used to extract the contour of the expanded image to obtain the contour of the landslide area.

[0081] In the step 7, the landslide area calculation module calculates the area of the forest landslide collapse, and the specific steps are as follows:

[0082] Step 71: The landslide area calculation module uses the `gdaltransform` tool in the open source raster spatial data conversion library GDAL to convert the pixel coordinates (row, column) in the remote sensing image into geographic coordinates (longitude, latitude); uses the `ogr2ogr` tool in the open source raster spatial data conversion library GDAL to save the converted landslide boundary geographic coordinates in vector format, such as Shapefile or GeoJSON, for further analysis and visualization in GIS software.

[0083] Step 72: Use the three-dimensional reconstruction open source software library OpenMVG to estimate camera pose and generate sparse point cloud, after generating sparse point cloud, further optimize the reconstruction model and generate dense point cloud, then use GDAL tool to convert the dense point cloud into digital elevation model DEM; Each point of the dense point cloud contains X, Y and Z coordinates, use `gdal_grid` command in GDAL to interpolate each point of the dense point cloud to generate regular DEM grid data;

[0084] Step 73: According to the DEM grid data, use the gdaldem command in the GDAL tool to calculate the slope; By creating a mask to extract the slope data of the landslide area in the remote sensing image, after extracting all the slope values in the landslide area, calculate the average value of all the slope values, the slope average value calculation formula is as follows:

[0085]

[0086] Where N is the total number of pixels in the landslide area, θ i is the slope value of the i-th pixel, unit is degree;

[0087] Step 74: According to the landslide area contour, get the coordinates (x i ,y i ) on the landslide area contour, then calculate the projection area of the landslide area on the horizontal plane, the calculation formula is as follows:

[0088]

[0089] Where (x i ,y i ) is the plane coordinates of the i-th vertex of the polygon, and n is the number of polygon vertices;

[0090] Then calculate the actual area of the landslide slope, that is, the forest landslide collapse area, using the average slope, the calculation formula is as follows:

[0091]

[0092] Where A 平面is the calculated planar projection area, and θ is the average slope of the landslide area.

[0093] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. The present application can be variously changed and modified by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for detecting forest landslide collapse area based on unmanned aerial vehicle (UAV) visual remote sensing, characterized in that, Comprise the following steps: Step 1: system construction: the construction is based on unmanned aerial vehicle vision remote sensing forest landslide collapse area detection system, the forest landslide collapse area detection system includes unmanned aerial vehicle, the unmanned aerial vehicle is carried with multispectral camera, the multispectral camera is connected in turn image preprocessing module, image alignment module, landslide area identification module and landslide area calculation module; Step 2: data acquisition: the unmanned aerial vehicle carries multispectral camera and carries out remote sensing image shooting on forest landslide area, obtains remote sensing image data, and transmits to image preprocessing module; Step 3: image preprocessing: the image preprocessing module obtains historical satellite image data of the forest landslide area, and carries out preprocessing operation on the historical satellite image data and remote sensing image data, obtains standard remote sensing image and historical image, and transmits to image alignment module; Step 4: image alignment: the image alignment module extracts feature points in the remote sensing image and historical image, and matches the feature points, then calculates the transformation matrix according to the matched feature points, and maps the remote sensing image and historical image to the same spatial reference system through the transformation matrix, to complete the alignment of images before and after landslide; Step 5: landslide area identification: the landslide area identification module identifies the landslide area by calculating the ratio vegetation index RVI of the remote sensing image and historical image; Step 6: image post-processing: the landslide area identification module uses the erosion and expansion operation of OpenCV to remove small noise and connect dispersed landslide area, and extracts the landslide area contour; Step 7: landslide area calculation: the landslide area calculation module uses GDAL tool to carry out coordinate conversion, generates digital elevation model DEM, calculates the average slope of the landslide area, and then combines the plane projection area and the average slope to calculate the actual area of the landslide inclined surface, to obtain the forest landslide collapse area.

2. The forest landslide collapse area detection method based on unmanned aerial vehicle vision remote sensing according to claim 1, characterized in that: In the step 2, the unmanned aerial vehicle carries multispectral camera to shoot multiple overlapping remote sensing images on the forest landslide area, and the heading overlap rate and the lateral overlap rate are not less than 80%. 3.The forest landslide collapse area detection method based on UAV visual remote sensing according to claim 1, characterized in that: In the step 3, the image preprocessing module carries out preprocessing operation on the remote sensing image data and historical satellite image data, and the specific steps are as follows: Step 31: radiation calibration: the image preprocessing module converts the remote sensing image data and historical satellite image data stored in the form of digital value DN into the form of radiation brightness value through conversion formula, and the conversion formula is as follows: L = (DN - DN offset ) x Gain where L is the radiance value, DN is the original digital value, DN offset is the offset value of the sensor, and Gain is the gain coefficient. Step 32: image correction: the image preprocessing module carries out processing operation including but not limited to geometric correction, image fusion, image cropping and shadow removal on the remote sensing image data and historical satellite image data in the form of radiation brightness value, and then uses FLAASH atmospheric correction model to correct the radiation reflectivity to ground reflectivity.

4. The forest landslide collapse area detection method based on unmanned aerial vehicle vision remote sensing according to claim 3, characterized in that: The geometric correction includes coarse correction, geometric fine correction and orthographic correction. 5.The forest landslide collapse area detection method based on UAV visual remote sensing according to claim 1, characterized in that: In the step 4, the image alignment module adopts SIFT scale invariant feature transform method to detect feature points in the remote sensing image and the historical image, and then matches the feature points in the remote sensing image and the historical image through a nearest neighbor search algorithm to obtain a feature point pair set. 6.The forest landslide collapse area detection method based on UAV visual remote sensing according to claim 1, characterized in that: In the step 5, the landslide area identification module reads the remote sensing image using an OpenCV (an open platform computer vision and machine learning software library), extracts data in red light bands and near-infrared bands, and calculates a ratio vegetation index (RVI) of the remote sensing image according to the following expression: where I RED is the red band data value of the corresponding pixel in the remote sensing image, I NIR is the near-infrared band data value of the corresponding pixel in the remote sensing image.

7. The forest landslide collapse area detection method based on unmanned aerial vehicle vision remote sensing according to claim 1, characterized in that: In the step 6, the landslide area identification module uses an erosion function in the OpenCV to perform erosion and expansion processing on the identified landslide area, and the specific steps are as follows: Step 61: a 3X3 erosion convolution kernel is selected, a cv2.erode() function in the OpenCV is used to perform erosion operation on the landslide area image to obtain an eroded image; Step 62: a 3X3 expansion convolution kernel is selected, a cv2.dilate() function in the OpenCV is used to perform expansion operation on the eroded image to obtain an expanded image; Step 63: a contour extraction function cv2.findContours() in the OpenCV is used to extract the contour of the expanded image to obtain a landslide area contour. 8.The forest landslide collapse area detection method based on UAV visual remote sensing according to claim 1 or 7, characterized in that: In the step 7, the landslide area calculation module calculates the forest landslide collapse area, and the specific steps are as follows: Step 71: the landslide area calculation module uses a `gdaltransform` tool in an open source raster spatial data conversion library GDAL to convert pixel coordinates in the remote sensing image into geographic coordinates, and uses an `ogr2ogr` tool in the open source raster spatial data conversion library GDAL to save the converted landslide boundary geographic coordinates in a vector format; Step 72: a camera pose estimation and sparse point cloud generation are performed using an open source software library OpenMVG for three-dimensional reconstruction, a reconstruction model is further optimized after the sparse point cloud is generated to generate a dense point cloud, and then a GDAL tool is used to convert the dense point cloud into a digital elevation model (DEM); each point of the dense point cloud contains X, Y and Z coordinates, and a `gdal_grid` command in the GDAL is used to interpolate each point of the dense point cloud to generate regular DEM grid data; Step 73: according to the DEM grid data, a gdaldem command in the GDAL tool is used to calculate the slope; a mask is created to extract the slope data of the landslide area in the remote sensing image, after all the slope values in the landslide area are extracted, an average value of all the slope values is calculated, and the average slope value calculation formula is as follows: where N is the total number of pixels in the landslide area, θ i is the slope value of the i-th pixel, in degrees. Step 74: According to the landslide area profile, the coordinates (x i ,y i ) on the landslide area profile are obtained, and then the projected area of the landslide area on the horizontal plane is calculated, and the calculation formula is as follows: where (x i ,y i ) is the planar coordinate of the i-th vertex of the polygon, and n is the number of vertices of the polygon. The average slope is used to calculate the actual area of the landslide slope, i.e., the forest landslide collapse area, and the calculation formula is as follows: where A 平面 is the calculated planar projection area, and θ is the average slope of the landslide area.

Images