Landslide area identification method, system and device and storage medium

Abstract

The invention provides a landslide area identification method, system and device and a storage medium, and belongs to the technical field of natural disaster risk management, and the method comprises the following steps: obtaining a video sequence image of a landslide monitoring area; modeling a background image of the video sequence image to obtain a background image in a state in which no landslide occurs; and taking the background image in the state of no landslide as a reference, extracting a change image in the video sequence image, judging whether the gray value of the change image reaches an early warning threshold value or not, if so, screening a graph with the largest contour in the change image, and identifying a landslide area. The method for extracting the landslide object based on the video image change features is high in universality and mobility, abnormal risk situations of water bodies in different areas can be rapidly recognized, and the early warning efficiency of the abnormal situations of the water bodies with risks is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of natural disaster risk management, and in particular relates to a landslide area identification method, system, equipment and storage medium. Background Art

[0002] If the signs and related characteristics of landslides can be identified in time and early warning notices can be issued so that relevant departments and personnel can take countermeasures earlier, the losses caused by landslides can be reduced.

[0003] At present, Wang Dong combines the landslide hazard zoning results with the real-time rainfall data collected by rainfall observation stations at key disaster points, and completes the regional landslide forecast and warning in the study area based on the geological-meteorological coupling warning model. Amrita Joshi et al. proposed a landslide prediction model based on WSN-AI, which predicts the landslide displacement and issues warnings based on ground sensor data and weather forecast data released by the meteorological department. This type of method requires real-time and accurate meteorological data, which can usually only be used internally by local meteorological bureaus and are difficult for other industries to obtain. Wang Wei used the BP neural network model to conduct landslide warning analysis for the Lishan landslide geological disaster, and built a monitoring data management platform based on Web GIS technology. However, the BP model training time is long and it is not suitable for dynamic real-time warning of sudden disasters. Wu Junnan proposed a double-layer LSTM landslide prediction model based on attention mechanism based on single-sequence landslide data, which solves the problem that the backward time information in the single-layer LSTM cannot be well utilized in historical data. However, the LSTM model training requires a large amount of long-series raw data, and the migration is poor, which is difficult to apply to river flood warning in different regions.

[0004] In summary, the difficulty in obtaining real-time and accurate meteorological data affects the training of landslide prediction models. At the same time, the long training time of existing landslide prediction models leads to low efficiency in landslide prediction. Summary of the invention

[0005] In order to overcome the low efficiency of landslide prediction in the above-mentioned prior art, the present invention provides a landslide area identification method, comprising the following steps:

[0006] Obtain video sequence images of the landslide monitoring area;

[0007] Modeling the background image of the video sequence image to obtain the background image in the state where no landslide occurs;

[0008] Taking the background image without landslide as the benchmark, the change image in the video sequence image is extracted, and it is judged whether the gray value of the change image reaches the warning threshold. If so, the figure with the largest contour in the change image is selected to identify the landslide area.

[0009] Preferably, the method of using a Gaussian mixture model to model the background image of the video sequence image to obtain the background image in a state where no landslide occurs comprises the following steps:

[0010] The gray value of each pixel in the video sequence image is described by the weighted sum of multiple Gaussian distributions, and a model is generated for each pixel to obtain multiple Gaussian distributions;

[0011] The first N Gaussian distributions are selected as the background model, and the video sequence images input into the Gaussian mixture model are matched with the first N Gaussian distributions. If the video sequence image matches successfully, it is considered to be the background image.

[0012] Preferably, if the video sequence image matches successfully, it is considered to be a background image, specifically: if the pixels of the input video sequence image are within 2.5 standard deviations of the Gaussian distribution, the match is successful.

[0013] Preferably, the method uses the background image in a state where no landslide occurs as a reference to extract the change image in the video sequence image, specifically: subtracting each frame image in the video sequence image from the background image to obtain a grayscale value difference, comparing the absolute value of the grayscale value difference between each frame image and the background image with a set threshold, and taking the image with an absolute value greater than the set threshold as the change image.

[0014] Preferably, before screening the graphics with the largest contour in the change image, the method further includes using an opening operation to eliminate noise and burrs in the change image.

[0015] The present invention also provides a landslide area identification system, comprising:

[0016] An image acquisition module is used to acquire video sequence images of the landslide monitoring area;

[0017] A background image acquisition module is used to model the background image of the video sequence image to obtain the background image in a state where no landslide occurs;

[0018] The landslide identification module is used to extract the change image in the video sequence image based on the background image in the state where no landslide has occurred, and to determine whether the grayscale value of the change image reaches the warning threshold. If so, the figure with the largest contour in the change image is screened to identify the landslide area.

[0019] The present invention also provides a computer device, comprising a memory and a processor; the memory stores a computer program, and the processor is used to run the computer program in the memory to execute the landslide area identification method.

[0020] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute the landslide area identification method.

[0021] The landslide area identification method provided by the present invention has the following beneficial effects:

[0022] The video sequence images of the landslide monitoring area of ​​the present invention do not rely on real-time high-precision meteorological data and are easy to obtain; by extracting the change images in the video sequence images, the landslide area can be identified according to the grayscale value of the change images, and in this process, a large amount of historical disaster data is not required for model training; the method of extracting landslide objects based on video image change features of the present invention is universal and has strong portability, and can quickly identify the abnormal risk situation of water bodies in different areas, thereby improving the efficiency of early warning of abnormal risk water body situations. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the embodiment of the present invention and its design scheme, the following briefly introduces the drawings required for this embodiment. The drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0024] Figure 1 A flow chart of a landslide area identification method according to an embodiment of the present invention;

[0025] Figure 2 Flowchart for extracting differential images using background subtraction method;

[0026] Figure 3 This is the result of sudden landslide disaster information extraction. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the technical solution of the present invention and implement it, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the scope of protection of the present invention.

[0028] In the description of the present invention, it is to be understood that the terms “center”, “longitudinal”, “lateral”, “length”, “width”, “thickness”, “up”, “down”, “front”, “back”, “left”, “right”, “vertical”, “horizontal”, “top”, “bottom”, “inside”, “outside”, “axial”, “radial”, “circumferential”, etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing the technical solutions of the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0029] In addition, the terms "first", "second", etc. are used for descriptive purposes only and are not to be understood as indicating or implying relative importance. In the description of the present invention, it should be noted that, unless otherwise clearly specified or limited, the terms "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances. In the description of the present invention, unless otherwise specified, "plurality" means two or more, which will not be described in detail here.

[0030] Example

[0031] The present invention provides a method for identifying a landslide area, specifically Figure 1 As shown, the following steps are included:

[0032] Step 1: Obtain video sequence images of the landslide monitoring area.

[0033] Step 2: Use the Gaussian mixture model to model the background image of the video sequence image to obtain the background image when no landslide occurs.

[0034] When monitoring and shooting pictures, the camera position is generally fixed, so it can be considered that the shooting scene is less changing. So the area can be divided into background and foreground, the background is generally considered to be static, and the foreground is considered to be changing. Gaussian mixture model background modeling is essentially a classification problem, specifically: the grayscale value of each pixel in the video sequence image is described by the weighted sum of multiple Gaussian distributions, a model is generated for each pixel, the first N Gaussian distributions are selected as the background model, and the input image is matched with the first N Gaussian distributions. If the match is successful, it is considered to be the background. The pixel is considered to be the background within 2.5 standard deviations of the Gaussian distribution, otherwise it is considered to be the foreground. If the match with N Gaussian distributions fails, the last Gaussian distribution is replaced. If it is considered to be the background, the Gaussian model is updated to generate a new Gaussian distribution. Therefore, the Gaussian mixture model background modeling method adopted by the present invention, although the method can detect the target object, there is more noise interference, and morphological processing is required.

[0035] Step 3: Taking the background image without landslide as the reference, extract the change image in the video sequence image and determine whether the gray value of the change image reaches the warning threshold. If so, select the graphic with the largest contour in the change image to identify the landslide area.

[0036] After establishing a background image, the background subtraction method is used to obtain the change image, that is, the difference image. The specific process is as follows Figure 2 As shown in the figure, the grayscale of each frame in the video sequence image is subtracted from the grayscale of the background image, a certain threshold is set, and the absolute value of the grayscale value difference is compared with the set threshold to determine the pixels of the moving target, and the image greater than the threshold is regarded as the changed image. The background subtraction method takes into account the continuous change of the background in the scene, and the background image will be continuously learned and updated according to the subsequent video frames.

[0037] The opening operation is used to eliminate the noise and burrs of the changing image, and then the graph with the largest contour in the changing image is screened out, which is the part that has changed. For a landslide, it is the part that has landslided, such as Figure 3 As shown, the landslide area is framed by a rectangle. Figure 3 Marked out in the figure is the part where the landslide is extracted. Figure 3 The risk index is the percentage of pixels in the changed image, which can reflect the degree of risk. The change index is the ratio of pixels changed relative to the baseline image, which can determine the degree of disaster occurrence.

[0038] The present invention also provides a landslide area identification system, comprising:

[0039] An image acquisition module is used to acquire video sequence images of the landslide monitoring area;

[0040] A background image acquisition module is used to model the background image of the video sequence image to obtain the background image in a state where no landslide occurs;

[0041] The landslide identification module is used to extract the change image in the video sequence image based on the background image in the state where no landslide has occurred, and to determine whether the grayscale value of the change image reaches the warning threshold. If so, the figure with the largest contour in the change image is screened to identify the landslide area.

[0042] The present invention also provides a computer device, comprising a memory and a processor; the memory stores a computer program, and the processor is used to run the computer program in the memory to execute the landslide area identification method.

[0043] The present invention also provides a computer-readable storage medium, which stores a computer program. The computer program is suitable for being loaded by a processor to execute the landslide area identification method.

[0044] The embodiments described above are only preferred specific implementation modes of the present invention, and the protection scope of the present invention is not limited thereto. Any simple changes or equivalent replacements of the technical solutions that can be obviously obtained by any technician familiar with the field within the technical scope disclosed in the present invention belong to the protection scope of the present invention.

Claims

1. A method for identifying landslide areas, characterized in that: The steps include: Obtain video sequence images of the landslide monitoring area; Modeling the background image of the video sequence image to obtain the background image in the state where no landslide occurs; Taking the background image without landslide as the benchmark, the change image in the video sequence image is extracted, and it is judged whether the gray value of the change image reaches the warning threshold. If so, the figure with the largest contour in the change image is selected to identify the landslide area.

2. The landslide area identification method according to claim 1, characterized in that: The method of modeling the background image of the video sequence image to obtain the background image in a state where no landslide occurs includes the following steps: The gray value of each pixel in the video sequence image is described by the weighted sum of multiple Gaussian distributions, and a model is generated for each pixel to obtain multiple Gaussian distributions; The first N Gaussian distributions are selected as the background model, and the video sequence images input into the Gaussian mixture model are matched with the first N Gaussian distributions. If the video sequence image matches successfully, it is considered to be the background image.

3. The landslide area identification method according to claim 2, characterized in that: If the video sequence image matches successfully, it is considered to be a background image. Specifically, if the pixels of the input video sequence image are within 2.5 standard deviations of the Gaussian distribution, the match is successful.

4. The landslide area identification method according to claim 1, characterized in that: The method uses the background image in a state where no landslide occurs as a reference to extract the change image in the video sequence image, specifically: subtract each frame image in the video sequence image from the background image to obtain a grayscale value difference, compares the absolute value of the grayscale value difference between each frame image and the background image with a set threshold, and takes the image with an absolute value greater than the set threshold as the change image.

5. The landslide area identification method according to claim 1, characterized in that: Before screening the graphics with the largest contour in the change image, the method further includes using an opening operation to eliminate noise and burrs in the change image.

6. A landslide area identification system, characterized in that: include: An image acquisition module is used to acquire video sequence images of the landslide monitoring area; A background image acquisition module is used to model the background image of the video sequence image using a Gaussian mixture model to obtain the background image in a state where no landslide occurs; The landslide identification module is used to extract the change image in the video sequence image based on the background image in the state where no landslide has occurred, and to determine whether the grayscale value of the change image reaches the warning threshold. If so, the figure with the largest contour in the change image is screened to identify the landslide area.

7. A computer device, characterized in that: It comprises a memory and a processor; the memory stores a computer program, and the processor is used to run the computer program in the memory to execute the landslide area identification method according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute the landslide area identification method according to any one of claims 1-5.

Images