Geological disaster monitoring and early warning system
Through adaptive partitioning and weighted fusion filtering algorithm, the problem of insufficient edge information retention in SAR images is solved, and the accuracy and image quality of geological disaster monitoring and early warning are improved.
Patent Information
- Application Number
- CN202411099723.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-12
- Publication Date
- 2024-12-03
AI Technical Summary
When the existing Lee filtering algorithm filters SAR images, it fails to effectively retain edge information, which affects the accuracy of geological disaster monitoring and early warning.
Adaptive partitioning algorithm is used to divide pixels with close gray values into the same temporary area, and filter pixels based on temporary areas, and combine Lee filtering algorithm and adaptive filtering algorithm to obtain the filtered gray through weighted summing. degree value to retain more edge information.
The filtered image quality is improved, the accuracy of geological disaster monitoring and early warning is enhanced, and the removal of spot noise and edge information is effectively preserved.
Smart Images

Figure CN119068329A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of disaster early warning, and in particular to a geological disaster monitoring and early warning system. Background Art
[0002] Using SAR images to warn of geological disasters such as landslides, debris flows, and ground collapse is a common warning method in the prior art. For example, patents such as CN117854253A, CN116343061A, and CN117705015A all disclose relevant warning technologies. When using SAR images for early warning, it is necessary to first filter the SAR images to improve the quality of the images used for identification. In the prior art, when filtering SAR images, the commonly used filtering algorithm is the Lee filtering algorithm. Lee algorithm filtering is a filtering method based on local statistical characteristics. It assumes that the noise in the image is multiplicative noise, and the noise intensity in the local area is linearly related to the image signal. The speckle noise in the SAR image is filtered by calculating the mean and variance of the pixel values in the local window.
[0003] However, the existing Lee filtering algorithm does not take into account the regional attributes of the pixels within the local window. When filtering the pixels located at the intersection of two different regions, it is easy to introduce too much grayscale information that does not belong to the region, resulting in the pixels near the edge being smoothed too much and the edge information being destroyed too much. The probability of occurrence is high, which affects the accuracy of the results obtained by subsequent identification, and thus affects the accuracy of geological disaster monitoring and early warning. Summary of the invention
[0004] The purpose of the present invention is to disclose a geological disaster monitoring and early warning system to solve the problem of how to retain more edge information for images filtered by the Lee algorithm when using SAR images for address disaster warning, thereby improving the quality of the filtered images and improving the accuracy of geological disaster monitoring and early warning.
[0005] In order to achieve the above object, the present invention adopts the following technical solution:
[0006] The present invention provides a geological disaster monitoring and early warning system, comprising a SAR image preprocessing device, the SAR image preprocessing device comprising a grayscale module, an adaptive partitioning module and a filtering module;
[0007] The grayscale module is used to convert the SAR image into a grayscale image;
[0008] The adaptive partitioning module is used to partition the pixels using an adaptive partitioning algorithm, and to divide the pixels with similar gray values into the same area to obtain multiple temporary areas;
[0009] The filtering module is used to filter each pixel in the grayscale image based on the temporary area to obtain a filtered SAR image;
[0010] Wherein, filtering each pixel in the grayscale image based on the temporary region includes:
[0011] S1, the pixel point to be filtered is represented as p;
[0012] S2, determine whether p is a pixel near the edge, if not, go to S3, if yes, go to S4;
[0013] S3, filter p using Lee filtering algorithm to obtain the gray value of p;
[0014] S4, filter p using the following method:
[0015] S41, calculate the filtering range distance d of p p ;
[0016] S42, the distance between p and p is less than d p The pixel points are stored in the collection pub;
[0017] S43, storing all pixel points in the temporary area where p is located into the set puc;
[0018] S44, obtain the intersection pud of pub and puc;
[0019] S45, use the following formula to obtain the gray value after p filtering:
[0020] fgray p =Leegray p,1 +(1-θ p )×Leegray p,2
[0021] fgray p Represents the gray value after p filtering, Leegray p,1 Represents the gray value obtained after filtering p using the Lee filtering algorithm based on the pixel points in pub, Leegray p,2 It represents the gray value obtained after filtering p based on the pixel points in pud using the adaptive filtering algorithm, θ p is the adaptive filter weight.
[0022] Preferably, it also includes a SAR image acquisition device, which is used to perform remote sensing on the monitoring area and acquire a SAR image of the monitoring area.
[0023] Preferably, it also includes a SAR image recognition device, which is used to recognize the filtered SAR image and obtain the geological disaster information existing in the filtered SAR image.
[0024] Preferably, the geological disaster information includes all types of geological disasters contained in the filtered SAR image and the distribution area of each type of geological disaster.
[0025] Preferably, it also includes a monitoring and early warning device,
[0026] The monitoring and early warning device is used to provide geological disaster early warning based on geological disaster information.
[0027] Preferably, an adaptive partitioning algorithm is used to partition the pixels, and pixels with similar grayscale values are divided into the same area to obtain multiple temporary areas, including:
[0028] S10, storing all pixels in the grayscale image into a set ALU;
[0029] S20, randomly extract a pixel r from ALU;
[0030] S30, performing adaptive partitioning processing on r to obtain a temporary area corresponding to r;
[0031] S40, deleting the pixel points in the temporary area corresponding to r from the ALU;
[0032] S50, determining whether there are still pixels in the ALU, if so, entering S20, if not, ending the partitioning process.
[0033] Preferably, performing adaptive partitioning processing on r to obtain a temporary area corresponding to r includes:
[0034] S100, taking r as a partition pixel point and storing r in a set temp;
[0035] S200, obtaining a set orcp of pixel points in an 8-neighborhood of a partitioned pixel point from the ALU;
[0036] S300, determining whether there are pixels in orcp that meet the set partitioning conditions, if so, selecting a pixel from the pixels that meet the set partitioning conditions as the next partitioning pixel, and entering S400; if not, entering S500;
[0037] S400, storing the new partitioned pixel points into the set temp, and entering S200;
[0038] S500: The area formed by the pixels in temp is used as the temporary area corresponding to r.
[0039] Preferably, judging whether there are pixels in the orcp that meet the set partitioning conditions includes:
[0040] Calculate the grayscale difference between each pixel in orcp and r respectively:
[0041] difgray i,r = | gray i -gray r |
[0042] gray r Represents the grayscale difference between pixel i and r in orcp, gray i and gray r Represent the grayscale values of pixel i and r respectively;
[0043] Determine whether there is a grayscale difference less than the set grayscale difference threshold. If so, it means that there are pixels in the orcp that meet the set partitioning conditions; if not, it means that there are no pixels in the orcp that meet the set partitioning conditions.
[0044] Preferably, selecting a pixel point as the next partition pixel point from the pixel points meeting the set partition condition includes:
[0045] Among all the pixels that meet the set partitioning conditions, the pixel with the smallest corresponding grayscale difference is used as the next partitioning pixel.
[0046] Preferably, determining whether p is a pixel point located near an edge includes:
[0047] Get the set of pixels nepxu in a window of size D×D centered on p;
[0048] Determine whether all the pixels in nepxu belong to the same temporary area. If so, it means that p does not belong to the pixel near the edge; if not, it means that p belongs to the pixel near the edge.
[0049] Beneficial effects:
[0050] The geological disaster monitoring and early warning system of the present invention introduces edge information obtained based on a temporary area during the filtering of SAR images, so that different filtering algorithms can be used based on different reference ranges to obtain the filtered grayscale value of the same pixel point, and then the two grayscale values are weighted and summed. While effectively filtering the speckle noise, more edge information can be retained, thereby improving the quality of the image obtained after filtering, and further effectively improving the accuracy of geological disaster monitoring and early warning. BRIEF DESCRIPTION OF THE DRAWINGS
[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.
[0052] Figure 1 The present invention is a schematic diagram of a geological disaster monitoring and early warning system.
[0053] Figure 2 Another schematic diagram of a geological disaster monitoring and early warning system of the present invention. DETAILED DESCRIPTION
[0054] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.
[0055] like Figure 1 In one embodiment shown, the present invention provides a geological disaster monitoring and early warning system, including a SAR image preprocessing device, the SAR image preprocessing device includes a grayscale module, an adaptive partitioning module and a filtering module;
[0056] The grayscale module is used to convert the SAR image into a grayscale image;
[0057] The adaptive partitioning module is used to partition the pixels using an adaptive partitioning algorithm, and to divide the pixels with similar gray values into the same area to obtain multiple temporary areas;
[0058] The filtering module is used to filter each pixel in the grayscale image based on the temporary area to obtain a filtered SAR image;
[0059] Wherein, filtering each pixel in the grayscale image based on the temporary region includes:
[0060] S1, the pixel point to be filtered is represented as p;
[0061] S2, determine whether p is a pixel near the edge, if not, go to S3, if yes, go to S4;
[0062] S3, filter p using Lee filtering algorithm to obtain the gray value of p;
[0063] S4, filter p using the following method:
[0064] S41, calculate the filtering range distance d of p p ;
[0065] S42, the distance between p and p is less than d p The pixel points are stored in the collection pub;
[0066] S43, storing all pixel points in the temporary area where p is located into the set puc;
[0067] S44, obtain the intersection pud of pub and puc;
[0068] S45, use the following formula to obtain the gray value after p filtering:
[0069] fgray p =Leegray p,1 +(1-θ p )×Leegray p,2
[0070] fgray p Represents the gray value after p filtering, Leegray p,1 Represents the gray value obtained after filtering p using the Lee filtering algorithm based on the pixel points in pub, Leegray p,2 It represents the gray value obtained after filtering p based on the pixel points in pud using the adaptive filtering algorithm, θ p is the adaptive filter weight.
[0071] The above filtering process is not only based on the traditional Lee algorithm for filtering, nor is it like the prior art, based on the same filtering range, using two different filtering algorithms for filtering, and then fusing the filtered grayscale values. Instead, it obtains another filtering range of different sizes containing more edge information, uses different filtering algorithms to filter the two filtering algorithms respectively, and finally fuses the filtering results. In this way, more edge information can be introduced into the pixel points during the filtering process, reducing the smoothness of the filtered pixel points, thereby retaining more edge information for the image obtained after filtering, and effectively improving the quality of the filtered image.
[0072] Preferably, if Figure 2As shown, it also includes a SAR image acquisition device, which is used to perform remote sensing on the monitoring area and acquire a SAR image of the monitoring area.
[0073] Specifically, the SAR image acquisition device may acquire SAR images of the same monitoring area at a fixed period, for example, remote sensing once every other week.
[0074] Preferably, the SAR image acquisition device comprises a SAR acquisition module, a signal processing module and an image generation module;
[0075] The SAR acquisition module acquires echo signals in the following way:
[0076] Radar emission: microwave signals are emitted to the surface of the earth. SAR usually uses different frequency bands such as C-band, X-band, and L-band to meet different observation needs.
[0077] Radar reception: When microwave signals encounter objects on the ground, part of the signal is reflected back to the radar receiving antenna. Due to the different physical properties of the ground (such as roughness, humidity, vegetation, etc.), the intensity and phase of the reflected signal will be different, and these differences are used to form the basic information of the image.
[0078] Pulse modulation: Use frequency modulated pulse (Chirp) signals to increase distance resolution. The frequency of the transmitted signal gradually increases or decreases within a pulse, so that the received signal can improve the resolution through pulse compression technology.
[0079] The signal processing module processes the echo signal in the following way:
[0080] Range compression: The received echo signal is first processed by pulse compression, which uses the modulation characteristics of the transmitted signal to compress the long pulse signal into short pulses to improve the range resolution.
[0081] Azimuth synthetic aperture: The azimuth resolution is improved through the "synthetic aperture" technology. That is, through the movement of the platform, the surface echo signals received by the radar antenna at different locations are combined, which is equivalent to generating a larger antenna (synthetic aperture). This method makes the azimuth resolution much higher than the resolution of the actual antenna size.
[0082] The image generation module is used to process the signal processed by the signal processing module in the following manner:
[0083] In the signal processing process, the compression results in the range and azimuth directions are used to convert the original echo data into a two-dimensional image through Fourier transform and other algorithms. This process includes range compression (Range Compression) and azimuth compression (Azimuth Compression), and finally generates a SAR image.
[0084] Preferably, the SAR image acquisition device further comprises a processing module, and the processing module is used to perform the following processing:
[0085] Image correction: The generated SAR images usually need to be geometrically corrected to eliminate geometric distortion caused by factors such as platform motion and terrain undulation.
[0086] Radiation correction: Perform radiation correction on SAR images to eliminate radiation non-uniformity caused by factors such as incident angle and surface reflection characteristics, ensuring the accuracy of the physical meaning of the image.
[0087] Preferably, it also includes a SAR image recognition device, which is used to recognize the filtered SAR image and obtain the geological disaster information existing in the filtered SAR image.
[0088] Specifically, any of the following algorithms can be used to obtain geological disaster information:
[0089] Support Vector Machine (SVM): By training the SVM model, the types of geological disasters can be classified based on features (such as texture, shape, and reflection intensity), such as identifying landslides and mud-rock flows.
[0090] Random Forest: Random Forest is an integrated learning method that can effectively classify complex disaster types through the combination of multiple decision trees.
[0091] k-means clustering: used to automatically segment and classify disaster areas, especially when the prior labels are uncertain, k-means can help identify and estimate the area of geological disasters.
[0092] Preferably, the geological disaster information includes all types of geological disasters contained in the filtered SAR image and the distribution area of each type of geological disaster.
[0093] Preferably, the types of geological disasters include landslides, subsidence, mud-rock flows, and the like.
[0094] The image recognition device is used to identify the types of all geological disasters existing in the filtered SAR image and the distribution area of each type of geological disaster when there are geological disasters in the filtered SAR image.
[0095] Preferably, it also includes a monitoring and early warning device,
[0096] The monitoring and early warning device is used to provide geological disaster early warning based on geological disaster information.
[0097] Specifically, geological disaster early warning is carried out based on geological disaster information, including:
[0098] Obtain the growth ratio of the distribution area of the same type of geological disaster in the same monitoring area in the two most recent remote sensing SAR images. When the growth ratio is greater than the set ratio threshold, send a text message to issue a geological disaster warning. The content of the text message includes the type of geological disaster and the distribution range of the geological disaster.
[0099] Different ratio thresholds can be set for different types of geological disasters. For example, for landslides, the ratio threshold can be set to 0.1.
[0100] The calculation method of the growth rate of distribution area is as follows: h and S h-1 represents the distribution area of the same type of geological disasters in the SAR images obtained by remote sensing at the hth and h-1th times, and the growth ratio is
[0101] Preferably, an adaptive partitioning algorithm is used to partition the pixels, and pixels with similar grayscale values are divided into the same area to obtain multiple temporary areas, including:
[0102] S10, storing all pixels in the grayscale image into a set ALU;
[0103] S20, randomly extract a pixel r from ALU;
[0104] S30, performing adaptive partitioning processing on r to obtain a temporary area corresponding to r;
[0105] S40, deleting the pixel points in the temporary area corresponding to r from the ALU;
[0106] S50, determining whether there are still pixels in the ALU, if so, entering S20, if not, ending the partitioning process.
[0107] In the process of obtaining the temporary area, each time after the temporary area to which a pixel belongs is determined, the pixel is deleted from the ALU, so that the number of pixels in the ALU can be gradually reduced as the partitioning process proceeds. At the same time, repeated partitioning can be avoided, which can improve the efficiency of partitioning.
[0108] Preferably, performing adaptive partitioning processing on r to obtain a temporary area corresponding to r includes:
[0109] S100, taking r as a partition pixel point and storing r in a set temp;
[0110] S200, obtaining a set orcp of pixel points in an 8-neighborhood of a partitioned pixel point from the ALU;
[0111] S300, determining whether there are pixels in orcp that meet the set partitioning conditions, if so, selecting a pixel from the pixels that meet the set partitioning conditions as the next partitioning pixel, and entering S400; if not, entering S500;
[0112] S400, storing the new partitioned pixel points into the set temp, and entering S200;
[0113] S500: The area formed by the pixels in temp is used as the temporary area corresponding to r.
[0114] During the process of adaptive partitioning, the present invention continuously incorporates pixel points with smaller gaps with partitioned pixel points into the set temp, and also updates the positions of the partitioned pixel points, so that the temporary area can be gradually expanded along the direction of the real image edge, and the gaps between the pixel points in the temporary area can be further reduced. At the same time, information closer to the real image edge can be introduced into the temporary area, so that the clarity of the edges in the filtered image can be improved during subsequent filtering.
[0115] Preferably, judging whether there are pixels in the orcp that meet the set partitioning conditions includes:
[0116] Calculate the grayscale difference between each pixel in orcp and r respectively:
[0117] difgray i,r = | gray i -gray r |
[0118] gray r Represents the grayscale difference between pixel i and r in orcp, gray i and gray r Represent the grayscale values of pixel i and r respectively;
[0119] Determine whether there is a grayscale difference less than the set grayscale difference threshold. If so, it means that there are pixels in the orcp that meet the set partitioning conditions; if not, it means that there are no pixels in the orcp that meet the set partitioning conditions.
[0120] Preferably, the grayscale difference threshold is set to gray mx and gray mi Respectively represent the maximum and minimum grayscale values of the grayscale image.
[0121] Preferably, selecting a pixel point as the next partition pixel point from the pixel points meeting the set partition condition includes:
[0122] Among all the pixels that meet the set partitioning conditions, the pixel with the smallest corresponding grayscale difference is used as the next partitioning pixel.
[0123] Preferably, determining whether p is a pixel point located near an edge includes:
[0124] Get the set of pixels nepxu in a window of size D×D centered on p;
[0125] Determine whether all the pixels in nepxu belong to the same temporary area. If so, it means that p does not belong to the pixel near the edge; if not, it means that p belongs to the pixel near the edge.
[0126] Due to the existence of image noise, if a traditional edge detection algorithm is used, the edge of the obtained image is less accurate because it has not been filtered, and noise pixels are easily identified as edge pixels. However, the present invention is different. The present invention adopts an adaptive partitioning method to obtain multiple temporary areas. In the process of obtaining temporary areas, noise with a large difference in grayscale value from normal pixels will be bypassed by the temporary area that does not contain noise, and will not be included in the temporary area where the normal pixels are located. Therefore, the noise content in most temporary areas can be effectively reduced, which is conducive to obtaining higher quality filtering results.
[0127] Preferably, the value of D is 7.
[0128] Preferably, the filter range distance d of p is calculated p ,include:
[0129]
[0130] d rs is the maximum value of the set filter range distance, ncp represents the set of pixels in a window of size 7×7 centered on p, and grds j represents the gradient value of pixel j, miagrds represents the average gradient value of the pixel points in ncp, Nnfp represents the number of pixels in ncp that are not in the same temporary area as p, G() represents the normalization of the values in brackets, w1 and w2 represent the first filter range distance parameter and the second filter range distance parameter, respectively.
[0131] The filtering range of the present invention is calculated based on the gradient value and the number of pixels not in the same temporary area. When the gradient value of the pixels around p is larger, the number of pixels not in the same temporary area is larger, that is, closer to the real image edge, the corresponding filtering range distance is smaller, so that in the area where the edge details of the image are densely distributed, a smaller range of pixels can be referenced for filtering to better retain the edge information of the image. When the gradient value of the pixels around p is smaller, the number of pixels not in the same temporary area is smaller, the filtering range distance is larger, so that more pixels are introduced for reference to obtain a more accurate filtering result.
[0132] Preferably, the values of the first filter range distance parameter and the second filter range distance parameter are respectively and
[0133] Preferably, d rs The value of is 10.
[0134] Preferably, based on the pixels in pub, p is filtered using a Lee filtering algorithm, including:
[0135] Use the following formula to calculate the gray value after p filtering:
[0136]
[0137] Lee Gray p Indicates the gray value of p after filtering p using the Lee filtering algorithm, aveGray pub Represents the mean gray value of the pixels in pub, gray p represents the gray value before p filtering, σ 2 represents the noise variance of the grayscale image, It represents the variance of the grayscale value of the pixel in pub, and λ represents the regularization parameter, which is used to control the smoothness of the filter.
[0138] Preferably, the regularization parameter is 0.12 or 0.8. λ is a small value, for example, between 0 and 1. A small λ value means that the filter is more inclined to preserve image details rather than over-smoothing. The value of λ is usually proportional to the noise variance of the image. If the noise variance is large, it may be necessary to increase the value of λ to enhance the filtering effect.
[0139] Preferably, based on the pixel points in pud, the gray value obtained after filtering p using an adaptive filtering algorithm includes:
[0140] Use the following formula to calculate the gray value after p filtering:
[0141]
[0142]
[0143] dapGray p It indicates the gray value obtained after filtering p using the adaptive filtering algorithm, dast k,p represents the distance between k and p, gray k Represents the gray value of k, gray p represents the gray value before p filtering, G() represents the normalization of the value in the brackets, g1 and g2 represent the first adaptive filter coefficient and the second adaptive filter coefficient respectively, and wei represents the gray difference coefficient.
[0144] In the process of adaptive filtering, the smaller the distance between pixel point k and pixel point p, the smaller the absolute value of the grayscale difference between pixel point k and pixel point p, the stronger the influence of the grayscale value of k on the grayscale value finally obtained by filtering during filtering. In this way, more image edge information can be retained. In addition, the pud and pub of the present invention are different. The pixels in pud belong to the same temporary area, and the gap between such pixels is small. After filtering, it does not introduce the grayscale information in another temporary area. In this way, excessive smoothing of the edge of the image can be avoided, and more image edge information can be retained, so that in S45, more real edge information is introduced for the grayscale value obtained by traditional Lee filtering to improve the quality of the filtered image.
[0145] Preferably, the first adaptive filter coefficient and the second adaptive filter coefficient are both
[0146] Preferably, the calculation formula of the adaptive filtering weight is:
[0147]
[0148] ars max represents the maximum area of the temporary region, nekp represents the set of all temporary regions adjacent to the temporary region where p is located, ars v represents the area of the temporary region v, Nnekp represents the total number of temporary regions in nekp, Θ is the filter weight control parameter, and G() indicates normalization of the values in brackets.
[0149] The value of the adaptive filtering weight of the present invention is related to the filtering range distance and the area of the temporary area. When the filtering range distance is smaller and the average area of the adjacent temporary area is smaller, the adaptive filtering weight is smaller. In this way, in S45, the gray value obtained after filtering p using the Lee filtering algorithm can be corrected with greater force, and more edge information can be introduced; when the filtering range distance is larger and the average area of the adjacent temporary area is larger, the correction force is reduced to avoid excessive correction and the introduction of erroneous edge information. Therefore, compared with the fixed weight, the adaptive filtering weight of the present invention can obtain a higher quality filtering result after the fusion of the two filtering algorithms.
[0150] Preferably, the value of the filter weight control parameter is
[0151] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.
Claims
1. A geological disaster monitoring and early warning system, characterized in that: It includes a SAR image preprocessing device, which includes a grayscale module, an adaptive partitioning module and a filtering module; The grayscale module is used to convert the SAR image into a grayscale image; The adaptive partitioning module is used to partition the pixels using an adaptive partitioning algorithm, and to divide the pixels with similar gray values into the same area to obtain multiple temporary areas; The filtering module is used to filter each pixel in the grayscale image based on the temporary area to obtain a filtered SAR image; Wherein, filtering each pixel in the grayscale image based on the temporary region includes: S1, the pixel point to be filtered is represented as p; S2, determine whether p is a pixel near the edge, if not, go to S3, if yes, go to S4; S3, filter p using Lee filtering algorithm to obtain the gray value of p; S4, filter p using the following method: S41, calculate the filtering range distance d of p p ; S42, the distance between p and p is less than d p The pixel points are stored in the collection pub; S43, storing all pixel points in the temporary area where p is located into the set puc; S44, obtain the intersection pud of pub and puc; S45, use the following formula to obtain the gray value after p filtering: fgray p =Leegray p,1 +(1-θ p )×Leegray p,2 fgray p Represents the gray value after p filtering, Leegray p,1 Represents the gray value obtained after filtering p using the Lee filtering algorithm based on the pixel points in pub, Leegray p,2 It represents the gray value obtained after filtering p based on the pixel points in pud using the adaptive filtering algorithm, θ p is the adaptive filter weight.
2. A geological disaster monitoring and early warning system according to claim 1, characterized in that: It also includes a SAR image acquisition device, which is used to perform remote sensing on the monitoring area and acquire a SAR image of the monitoring area.
3. A geological disaster monitoring and early warning system according to claim 1, characterized in that: It also includes a SAR image recognition device, which is used to recognize the filtered SAR image and obtain the geological disaster information existing in the filtered SAR image.
4. A geological disaster monitoring and early warning system according to claim 3, characterized in that: The geological disaster information includes all the types of geological disasters contained in the filtered SAR images and the distribution area of each type of geological disasters.
5. A geological disaster monitoring and early warning system according to claim 3, characterized in that: It also includes monitoring and early warning devices, The monitoring and early warning device is used to provide geological disaster early warning based on geological disaster information.
6. A geological disaster monitoring and early warning system according to claim 1, characterized in that: Adopt adaptive partitioning algorithm to partition pixels, divide pixels with similar gray values into the same area, and obtain multiple temporary areas, including: S10, storing all pixels in the grayscale image into a set ALU; S20, randomly extract a pixel r from ALU; S30, performing adaptive partitioning processing on r to obtain a temporary area corresponding to r; S40, deleting the pixel points in the temporary area corresponding to r from the ALU; S50, determining whether there are still pixels in the ALU, if so, entering S20, if not, ending the partitioning process.
7. A geological disaster monitoring and early warning system according to claim 6, characterized in that: Adaptively partition r to obtain the temporary area corresponding to r, including: S100, taking r as a partition pixel point and storing r in a set temp; S200, obtaining a set orcp of pixel points in an 8-neighborhood of a partitioned pixel point from the ALU; S300, determining whether there are pixels in orcp that meet the set partitioning conditions, if so, selecting a pixel from the pixels that meet the set partitioning conditions as the next partitioning pixel, and entering S400; if not, entering S500; S400, storing the new partitioned pixel points into the set temp, and entering S200; S500: The area formed by the pixels in temp is used as the temporary area corresponding to r.
8. A geological disaster monitoring and early warning system according to claim 7, characterized in that: Determine whether there are pixels in orcp that meet the set partitioning conditions, including: Calculate the grayscale difference between each pixel in orcp and r respectively: difgray i,r =|gray i -gray r | gray r Represents the grayscale difference between pixel i and r in orcp, gray i and gray r Represent the grayscale values of pixel i and r respectively; Determine whether there is a grayscale difference less than the set grayscale difference threshold. If so, it means that there are pixels in the orcp that meet the set partitioning conditions; if not, it means that there are no pixels in the orcp that meet the set partitioning conditions.
9. A geological disaster monitoring and early warning system according to claim 8, characterized in that: Select a pixel point from the pixels that meet the set partitioning conditions as the next partitioning pixel point, including: Among all the pixels that meet the set partitioning conditions, the pixel with the smallest corresponding grayscale difference is used as the next partitioning pixel.
10. A geological disaster monitoring and early warning system according to claim 1, characterized in that: Determine whether p is a pixel near the edge, including: Get the set of pixels nepxu in a window of size D×D centered on p; Determine whether all the pixels in nepxu belong to the same temporary area. If so, it means that p does not belong to the pixel near the edge; if not, it means that p belongs to the pixel near the edge.
Citation Information
Patent Citations
Geological disaster recognition and early warning method and system based on remote sensing technology
CN116343061A
Early warning method, device, equipment and medium for geological disasters in mining area
CN117705015A
Remote mountainous area geological disaster monitoring and early warning system based on satellite communication technology
CN117854253A
Cited By
Geological disaster detection method based on image processing
CN120355703A
