Geological disaster early warning method based on artificial intelligence
Through the artificial intelligence-based landslide monitoring method, image data is used to screen early warning samples and perceive hash values to judge landslides, solving the problem of expensive equipment and low sensitivity of traditional methods and equipment, and achieving high-precision landslide monitoring and early warning.
Patent Information
- Application Number
- CN202411498031.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-25
- Publication Date
- 2024-11-22
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional landslide monitoring methods and equipment are expensive and are not sensitive enough to the slight displacement changes of the slope, making it difficult to effectively monitor and early warning.
Using a geological disaster warning method based on artificial intelligence, we use a camera to acquire standard image data on the slope, collect real-time image data, filter target warning samples, and use pixel constraint model and perceived hash value to determine whether a landslide has occurred.
It realizes low-cost and high-sensitivity landslide monitoring, improves early warning accuracy and response, and effectively avoids the problems of expensive equipment and low sensitivity of traditional methods.
Smart Images

Figure CN119007420A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of image processing, and in particular to a geological disaster early warning method based on artificial intelligence. Background Art
[0002] Landslide is a common natural disaster that poses a serious threat to people's lives and property safety. Due to its special geological conditions, slope landslides are difficult to monitor and warn. Traditional landslide monitoring methods, such as total station measurement and high-precision GPS monitoring, can provide accurate displacement data, but the equipment is expensive and complicated to operate, and it is not sensitive enough to small displacement changes of the slope. Therefore, it is particularly important to develop a low-cost and highly sensitive landslide monitoring method. Summary of the invention
[0003] In order to solve the above problems, the present invention proposes a geological disaster early warning method based on artificial intelligence.
[0004] The technical solution of the present invention is: a geological disaster early warning method based on artificial intelligence comprises the following steps:
[0005] S1, obtaining standard image data of the slope, and using a camera to collect real-time image data of the slope;
[0006] S2, based on the detection samples of standard image data, using the classifier to screen target warning samples from real-time image data;
[0007] S3. Determine whether a landslide occurs on the slope based on the target warning sample of the real-time image data.
[0008] The images collected when there is no landslide risk on the slope are used as standard image data.
[0009] Furthermore, S2 includes the following sub-steps:
[0010] S21, generating a detection sample using standard image data;
[0011] S22, inputting each pixel point of the detection sample and the real-time image data into the classifier to obtain the classification value of each pixel point;
[0012] S23. Filter target warning samples according to the classification value of each pixel in the real-time image data.
[0013] The beneficial effect of the above further scheme is: in the present invention, based on standard image data, some target warning samples are extracted from the real-time image data as the key warning range, and the pixel values of the real-time image data are traversed to eliminate normal pixels. In subsequent steps, only the target warning samples are analyzed and processed to reduce the image processing process.
[0014] Furthermore, S21 includes the following sub-steps:
[0015] S211, inputting the pixel point where the center of the standard image data and each pixel point into the pixel constraint model, and obtaining the optimal pixel position by Gaussian iteration;
[0016] S212, calculating the difference between the pixel value of each pixel point in the standard image data and the pixel value of the optimal pixel position, and taking the pixel values of the pixel points whose difference is greater than 0 as detection samples.
[0017] The beneficial effect of the above further scheme is: in the present invention, the pixel points of the standard image data are brought into the pixel constraint model for iterative processing to obtain the optimal pixel value position, and then the difference between the pixel value of the optimal pixel position and the pixel values of the remaining pixels is used to extract the detection sample. The detection sample contains a wide range of pixels, which can provide a basis for screening target warning samples for real-time image data.
[0018] Furthermore, in S211, the pixel constraint model The expression is:
[0019] ;
[0020] In the formula, Indicates the position of the pixel in the standard image data, Represents the horizontal coordinate of the pixel point in the standard image data, Represents the vertical coordinate of the pixel in the standard image data, Indicates the position of the pixel at the center of the standard image data. represents the Euclidean distance operation, represents the index, Represents the pixel value of a pixel in the standard image data, Represents the pixel value of the pixel at the center of the standard image data. Indicates the length of standard image data. Indicates the width of standard image data.
[0021] Furthermore, in S22, the classification value of the pixel in the real-time image data The expression is:
[0022] ;
[0023] In the formula, Represents the first element of the detection sample, Indicates the number of samples tested elements, Indicates the number of samples tested elements, Indicates the number of elements of the detection sample, Represents a symbolic function.
[0024] For any real number ,like ,but ;like ,but ;like ,but In simple terms, the sign function determines the positive or negative value of the input value.
[0025] Further, in S23, if the horizontal coordinate in the real-time image data is And the vertical axis is The pixels satisfy , it belongs to the target warning sample, otherwise it does not belong to the target warning sample; among them, The horizontal axis is And the vertical axis is The pixel value of the pixel point, The horizontal axis is And the vertical axis is The pixel value of the pixel point, The horizontal axis is And the vertical axis is The pixel value of the pixel point, The horizontal axis is And the vertical axis is The pixel value of the pixel point, The horizontal axis is And the vertical axis is The pixel value of the pixel.
[0026] Furthermore, S3 includes the following sub-steps:
[0027] S31, according to the position of each pixel point in the target warning sample of the real-time image data, extract the pixel points at the same position in the standard image data as the target comparison sample;
[0028] S32, respectively calculating the perceptual hash value of the target warning sample and the perceptual hash value of the target comparison sample;
[0029] S33, judging whether the perceived hash value of the target warning sample and the perceived hash value of the target comparison sample meet the warning conditions, if so, the slope has not landslided, otherwise, the slope has landslided.
[0030] The beneficial effect of the above further solution is that in the present invention, when calculating the perceptual hash, the image is discrete cosine transformed, the low-frequency coefficients are extracted, the low-frequency coefficients are quantized, and mapped into a binary form. Therefore, the perceptual hash value contains several 0 or 1 values.
[0031] If the warning conditions are met between the perceptual hash of the target warning sample and the perceptual hash of the target comparison sample, it means that the two samples have visually similar images, and the difference between the real-time image data and the standard image data is not large, and no landslide has occurred.
[0032] Furthermore, in S33, the expression of the warning condition is:
[0033] ;
[0034] In the formula, Represents the sum of all elements in the perceptual hash value of the target warning sample, Represents the sum of all elements in the perceptual hash value of the target comparison sample, Indicates the number of pixels in the target warning sample.
[0035] The beneficial effects of the present invention are as follows: the present invention discloses a geological disaster early warning method based on artificial intelligence, which utilizes detection samples of standard image data to extract target early warning samples from real-time image data, and then performs hash processing on the target early warning samples and target comparison samples to obtain the final disaster early warning result; the present invention effectively avoids the problems of expensive equipment and low sensitivity brought about by traditional landslide geological disaster early warning, analyzes and warns real-time images of slopes, effectively improves early warning accuracy, and enhances the response level of geological disaster early warning. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] Figure 1 Flowchart of the geological disaster early warning method based on artificial intelligence. DETAILED DESCRIPTION
[0037] The embodiments of the present invention will be further described below in conjunction with the accompanying drawings.
[0038] like Figure 1 As shown, the present invention provides a geological disaster early warning method based on artificial intelligence, comprising the following steps:
[0039] S1, obtaining standard image data of the slope, and using a camera to collect real-time image data of the slope;
[0040] S2, based on the detection samples of standard image data, using the classifier to screen target warning samples from real-time image data;
[0041] S3. Determine whether a landslide occurs on the slope based on the target warning sample of the real-time image data.
[0042] The images collected when there is no landslide risk on the slope are used as standard image data.
[0043] In this embodiment of the present invention, S2 includes the following sub-steps:
[0044] S21, generating a detection sample using standard image data;
[0045] S22, inputting each pixel point of the detection sample and the real-time image data into the classifier to obtain the classification value of each pixel point;
[0046] S23. Filter target warning samples according to the classification value of each pixel in the real-time image data.
[0047] In the present invention, based on standard image data, some target warning samples are extracted from real-time image data as the key warning range. The pixel values of the real-time image data are traversed to eliminate normal pixels. In subsequent steps, only the target warning samples are analyzed and processed to reduce the image processing flow.
[0048] In this embodiment of the present invention, S21 includes the following sub-steps:
[0049] S211, inputting the pixel point where the center of the standard image data and each pixel point into the pixel constraint model, and obtaining the optimal pixel position by Gaussian iteration;
[0050] S212, calculating the difference between the pixel value of each pixel point in the standard image data and the pixel value of the optimal pixel position, and taking the pixel values of the pixel points whose difference is greater than 0 as detection samples.
[0051] In the present invention, the pixel points of the standard image data are brought into the pixel constraint model for iterative processing to obtain the optimal pixel value position, and then the difference between the pixel value of the optimal pixel position and the pixel values of the remaining pixels is used to extract the detection sample. The detection sample contains a wide range of pixels, which can provide a basis for screening target warning samples for real-time image data.
[0052] In the embodiment of the present invention, in S211, the pixel constraint model The expression is:
[0053] ;
[0054] In the formula, Indicates the position of the pixel in the standard image data, Represents the horizontal coordinate of the pixel point in the standard image data, Represents the vertical coordinate of the pixel in the standard image data, Indicates the position of the pixel at the center of the standard image data. represents the Euclidean distance operation, represents the index, Represents the pixel value of a pixel in the standard image data, Represents the pixel value of the pixel at the center of the standard image data. Indicates the length of standard image data. Indicates the width of standard image data.
[0055] In the embodiment of the present invention, in S22, the classification value of the pixel point in the real-time image data The expression is:
[0056] ;
[0057] In the formula, Represents the first element of the detection sample, Indicates the number of samples tested elements, Indicates the number of samples tested elements, Indicates the number of elements of the detection sample, Represents a symbolic function.
[0058] For any real number ,like ,but ;like ,but ;like ,but In simple terms, the sign function determines the positive or negative value of the input value.
[0059] In the embodiment of the present invention, in S23, if the horizontal coordinate in the real-time image data is And the vertical axis is The pixels satisfy , it belongs to the target warning sample, otherwise it does not belong to the target warning sample; among them, The horizontal axis is And the vertical axis is The pixel value of the pixel point, The horizontal axis is And the vertical axis is The pixel value of the pixel point, The horizontal axis is And the vertical axis is The pixel value of the pixel point, The horizontal axis is And the vertical axis is The pixel value of the pixel point, The horizontal axis is And the vertical axis is The pixel value of the pixel.
[0060] In this embodiment of the present invention, S3 includes the following sub-steps:
[0061] S31, according to the position of each pixel point in the target warning sample of the real-time image data, extract the pixel points at the same position in the standard image data as the target comparison sample;
[0062] S32, respectively calculating the perceptual hash value of the target warning sample and the perceptual hash value of the target comparison sample;
[0063] S33, judging whether the perceived hash value of the target warning sample and the perceived hash value of the target comparison sample meet the warning conditions, if so, the slope has not landslided, otherwise, the slope has landslided.
[0064] In the present invention, when calculating the perceptual hash, the image is subjected to discrete cosine transform, low-frequency coefficients are extracted, the low-frequency coefficients are quantized, and mapped to binary form. Therefore, the perceptual hash value contains several 0 or 1 values.
[0065] If the warning conditions are met between the perceptual hash of the target warning sample and the perceptual hash of the target comparison sample, it means that the two samples have visually similar images, and the difference between the real-time image data and the standard image data is not large, and no landslide has occurred.
[0066] In the embodiment of the present invention, in S33, the expression of the warning condition is:
[0067] ;
[0068] In the formula, Represents the sum of all elements in the perceptual hash value of the target warning sample, Represents the sum of all elements in the perceptual hash value of the target comparison sample, Indicates the number of pixels in the target warning sample.
[0069] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.
Claims
1. A geological disaster early warning method based on artificial intelligence, characterized in that: The following steps are involved: S1, obtaining standard image data of the slope, and using a camera to collect real-time image data of the slope; S2, based on the detection samples of standard image data, using the classifier to screen target warning samples from real-time image data; S3. Determine whether a landslide occurs on the slope based on the target warning sample of the real-time image data.
2. The artificial intelligence-based geological disaster early warning method according to claim 1 is characterized in that: The S2 comprises the following sub-steps: S21, generating a detection sample using standard image data; S22, inputting each pixel point of the detection sample and the real-time image data into the classifier to obtain the classification value of each pixel point; S23. Filter target warning samples according to the classification value of each pixel in the real-time image data.
3. The artificial intelligence-based geological disaster early warning method according to claim 2 is characterized in that: The S21 comprises the following sub-steps: S211, inputting the pixel point where the center of the standard image data and each pixel point into the pixel constraint model, and obtaining the optimal pixel position by Gaussian iteration; S212, calculating the difference between the pixel value of each pixel point in the standard image data and the pixel value of the optimal pixel position, and taking the pixel values of the pixel points whose difference is greater than 0 as detection samples.
4. The artificial intelligence-based geological disaster early warning method according to claim 3 is characterized in that: In S211, the pixel constraint model The expression is: ; In the formula, Indicates the position of the pixel in the standard image data, Represents the horizontal coordinate of the pixel point in the standard image data, Represents the vertical coordinate of the pixel in the standard image data, Indicates the position of the pixel at the center of the standard image data. represents the Euclidean distance operation, represents the index, Represents the pixel value of a pixel in the standard image data, Represents the pixel value of the pixel at the center of the standard image data. Indicates the length of standard image data. Indicates the width of standard image data.
5. The artificial intelligence-based geological disaster early warning method according to claim 2 is characterized in that: In S22, the classification value of the pixel in the real-time image data The expression is: ; In the formula, Represents the first element of the detection sample, Indicates the number of samples tested elements, Indicates the number of samples tested elements, Indicates the number of elements of the detection sample, Represents a symbolic function.
6. The artificial intelligence-based geological disaster early warning method according to claim 2 is characterized in that: In S23, if the horizontal coordinate in the real-time image data is And the vertical coordinate is The pixels satisfy , it belongs to the target warning sample, otherwise it does not belong to the target warning sample; among them, The horizontal axis is And the vertical axis is The pixel value of the pixel point, The horizontal axis is And the vertical coordinate is The pixel value of the pixel point, The horizontal axis is And the vertical coordinate is The pixel value of the pixel point, The horizontal axis is And the vertical axis is The pixel value of the pixel point, The horizontal axis is And the vertical coordinate is The pixel value of the pixel.
7. The artificial intelligence-based geological disaster early warning method according to claim 1 is characterized in that: The S3 comprises the following sub-steps: S31, according to the position of each pixel point in the target warning sample of the real-time image data, extract the pixel points at the same position in the standard image data as the target comparison sample; S32, respectively calculating the perceptual hash value of the target warning sample and the perceptual hash value of the target comparison sample; S33, judging whether the perceived hash value of the target warning sample and the perceived hash value of the target comparison sample meet the warning conditions, if so, the slope has not landslided, otherwise, the slope has landslided.
8. The artificial intelligence-based geological disaster early warning method according to claim 7 is characterized in that: In S33, the expression of the warning condition is: ; In the formula, Represents the sum of all elements in the perceptual hash value of the target warning sample, Represents the sum of all elements in the perceptual hash value of the target comparison sample, Indicates the number of pixels in the target warning sample.
Citation Information
Patent Citations
Monitoring data processing method for earthquake prediction, earthquake prediction method and system
CN109085641A
Slope disaster identification system based on monitoring images
CN112861617A
Early warning method and device for mountain slope collapse, equipment, and readable storage medium
CN114021487A
Geological disaster monitoring and early warning system based on video recognition technology
CN114267155A
Intelligent monitoring and early warning system for side slope
CN116386302A
