A deformation monitoring method, system and device for urban road engineering

Abstract

The present application relates to the technical field of photogrammetry, and specifically to a deformation monitoring method, system and device for urban road engineering, which specifically include: calculating the first possibility that each edge pixel is a noise pixel through the similarity between feature points in any two urban road SAR images and the matching degree of edge pixels in their corresponding two edge images; calculating the second possibility that each edge pixel is a noise pixel based on the irregular distribution characteristics of the noise pixels; correcting the grayscale value of the edge pixel based on the above characteristics and the grayscale value of the edge pixel before and after denoising, and enhancing the denoised image based on the corrected grayscale value, thereby improving the denoising effect of the denoising algorithm; based on the enhanced denoised SAR image of the urban road, combining with the surface deformation detection algorithm to perform deformation monitoring of the urban road, thereby improving the accuracy and reliability of the deformation monitoring results of the urban road.

Description

Technical Field

[0001] The present application relates to the technical field of photogrammetry, and in particular to a deformation monitoring method, system and device for urban road engineering. Background Art

[0002] As cities continue to expand, urban road networks become increasingly complex, and the health of roads has an increasingly prominent impact on the smoothness and safety of urban traffic. Deformation monitoring of urban road projects can detect road diseases in a timely manner and prevent serious safety accidents such as road collapse. The demand for full life cycle deformation monitoring in the design, construction and operation and maintenance of urban infrastructure is growing. Through deformation monitoring, the health of urban roads can be evaluated, providing a scientific basis for road maintenance and renovation.

[0003] The monitoring effect of synthetic aperture radar differential interferometry technology (PS-InSAR) depends on the quality of SAR images. SAR images will produce speckle noise when they are acquired. When the existing Lee denoising algorithm is used for denoising, the edge information of the image will be lost during the filtering process, affecting the sharpness and details of the image. Therefore, it is necessary to correct the SAR image after Lee denoising to improve the image quality and improve the accuracy of urban road deformation monitoring. Summary of the invention

[0004] In order to solve the above technical problems, the purpose of this application is to provide a deformation monitoring method, system and device for urban road engineering. The technical solutions adopted are as follows:

[0005] In a first aspect, an embodiment of the present application provides a deformation monitoring method for an urban road project, the method comprising the following steps:

[0006] Collect various SAR images of urban roads;

[0007] Taking the neighborhood of each feature point in each SAR image as each pixel block, and constructing the similarity between any two SAR images based on the gray value difference between the matched pixel blocks in any two SAR images;

[0008] Acquire each pair of matched edge pixel points between the edge images of the arbitrary two SAR images and the similarity between the matched edge pixel points through a matching algorithm; calculate a first possibility that each edge pixel point is a noise pixel point based on the similarity and the matching condition of each edge pixel point in each edge image and in combination with the similarity between the arbitrary two SAR images;

[0009] Determine a second possibility that each edge pixel is a noise pixel based on the length of the edge line where each edge pixel is located and the number of edge pixels in the neighborhood in the edge image, and calculate the noise level of each edge pixel in combination with the first possibility;

[0010] Calculating a final grayscale value of each edge pixel based on the noise level of each edge pixel and the grayscale values ​​of the corresponding pixel in the SAR image before and after denoising, and enhancing the denoised SAR image based on the final grayscale value;

[0011] Based on the enhanced denoised images of all SAR images, urban road deformation monitoring is carried out in combination with the surface deformation detection algorithm.

[0012] In one embodiment, the process of obtaining the similarity between any two SAR images is as follows:

[0013] Matching the pixel blocks of any two SAR images by a block matching algorithm to obtain matching pixel blocks of each pixel block in one SAR image in the other SAR image;

[0014] The gray value difference between the central pixel points of the two matched pixel blocks is recorded as a first difference; the mean of all the first differences between the arbitrary two SAR images is calculated, and the reciprocal of the normalized value of the mean is recorded as the similarity between the arbitrary two SAR images.

[0015] In one embodiment, the process of obtaining the first possibility that each edge pixel is a noise pixel is as follows:

[0016] If the current edge pixel does not have a matching edge pixel in a certain edge image, the certain edge image is used as the unmatched image of the current edge pixel; the first possibility that the uth edge pixel in the ath edge image is a noise pixel is recorded as , The expression is:

[0017]

[0018] In the formula, is the normalization function, represents the number of unmatched images of the u-th edge pixel in the a-th edge image, S represents the number of collected SAR images, represents the normalized value of the similarity between the a-th and b-th SAR images, Represents the similarity between the u-th edge pixel in the a-th edge image and the matching edge pixel in the b-th edge image.

[0019] In one embodiment, the process of obtaining the second possibility that each edge pixel is a noise pixel is as follows:

[0020] The straight lines in each edge image are obtained by a straight line detection algorithm, and the edge pixel points on the straight line whose length is greater than a preset length threshold are recorded as road edge pixel points; the second possibility that the road edge pixel point is a noise pixel point is set to 0;

[0021] For edge pixels other than road edge pixels, the number of edge pixels that are not collinear with the target edge pixel in the neighborhood of the target edge pixel is counted, recorded as the first number, the length of the edge line where the target edge pixel is located is obtained, recorded as the first length, and the second possibility of each edge pixel being a noise pixel is determined based on the first number and the first length, wherein the second possibility of each edge pixel is negatively correlated with the first number and the first length of each edge pixel, respectively.

[0022] In one embodiment, the noise level of each edge pixel is: the average of the first possibility and the second possibility of each edge pixel.

[0023] In one embodiment, the final grayscale value of each edge pixel is expressed as:

[0024] , where Represents the final grayscale value of the current edge pixel. and Respectively represent the grayscale value of the pixel in the SAR image before and after denoising corresponding to the current edge pixel, Indicates the noise level of the current edge pixel.

[0025] In one embodiment, the method of enhancing the denoised SAR image is: using the final grayscale value of each edge pixel to replace the grayscale value of the corresponding pixel in the denoised SAR image.

[0026] In one embodiment, the urban road deformation monitoring is specifically:

[0027] The enhanced and denoised images of all SAR images are used as the input of the surface deformation detection algorithm, and the output is the deformation variable of each point on the urban road; if the deformation variable of any point in the urban road is greater than the deformation threshold, an alarm is performed; otherwise, no alarm is performed.

[0028] In a second aspect, the embodiment of the present application further provides a deformation monitoring system for urban road engineering, including:

[0029] Data acquisition module: collects various SAR images of urban roads;

[0030] Image similarity analysis module: taking the neighborhood of each feature point in each SAR image as each pixel block, and constructing the similarity between any two SAR images based on the gray value difference between the matching pixel blocks in any two SAR images;

[0031] Noise possibility assessment module: obtains each pair of matched edge pixel points between the edge images of any two SAR images and the similarity between the matched edge pixel points through a matching algorithm; calculates the first possibility that each edge pixel point is a noise pixel point based on the similarity and the matching situation of each edge pixel point in each edge image and the similarity between the any two SAR images;

[0032] Noise degree calculation module: determines the second possibility that each edge pixel is a noise pixel based on the length of the edge line where each edge pixel is located in the edge image and the number of edge pixels in the neighborhood, and calculates the noise degree of each edge pixel in combination with the first possibility;

[0033] Image enhancement module: calculates the final grayscale value of each edge pixel based on the noise level of each edge pixel and the grayscale value of the corresponding pixel in the SAR image before and after denoising, and enhances the denoised SAR image based on the final grayscale value;

[0034] Urban road deformation monitoring module: Based on the enhanced denoised images of all SAR images, urban road deformation monitoring is performed in combination with the surface deformation detection algorithm.

[0035] In the third aspect, an embodiment of the present application also provides a deformation monitoring device for an urban road project, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the method described in the first aspect when executing the computer program.

[0036] The embodiments of the present application have at least the following beneficial effects:

[0037] The present application calculates the similarity between any two urban road SAR images through the similarity between the feature points in any two urban road SAR images; for the edge images of all urban road SAR images, the first possibility that each edge pixel is a noise pixel is calculated based on the matching degree of edge pixels in any two edge images and the similarity between the corresponding two urban road SAR images; the second possibility that each edge pixel is a noise pixel is calculated based on the irregular distribution characteristics of the noise pixels; the grayscale value of the edge pixel is corrected based on the above characteristics and the grayscale value of the edge pixel before and after denoising, and the image filtered by the Lee algorithm is enhanced based on the corrected grayscale value, so that the filtered image can maintain its denoising effect and greatly retain its edge information, which greatly increases the denoising effect of the Lee algorithm; based on the enhanced denoised SAR image of the urban road, the urban road deformation monitoring is carried out in combination with the surface deformation detection algorithm, thereby improving the accuracy and reliability of the urban road deformation monitoring results. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0039] Figure 1 A flow chart of the steps of a deformation monitoring method for an urban road project provided in one embodiment of the present application;

[0040] Figure 2 The figure is a structural schematic diagram of a deformation monitoring system for urban road engineering. DETAILED DESCRIPTION

[0041] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following is a detailed description of the deformation monitoring method, system and device for urban road engineering proposed in the present application, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0042] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0043] The following is a detailed description of a deformation monitoring method, system and device for urban road engineering provided by the present application in conjunction with the accompanying drawings.

[0044] See also Figure 1 , which shows a flow chart of a deformation monitoring method for an urban road project provided by an embodiment of the present application, the method comprising the following steps:

[0045] Step S1, collecting SAR images of urban roads.

[0046] When monitoring road deformation, the SAR images of the urban roads to be monitored are obtained through the Sentinel-1A satellite.

[0047] Preferably, in one embodiment of the present application, when the satellite collects SAR images of urban roads, the satellite working mode is set to IW, the image width is set to 200 km, the polarization mode adopts VV polarization, and the image collection interval is 30 days, that is, a SAR image is obtained every 30 days, and a total of 31 images are collected by Sentinel-1A orbit raising.

[0048] Step S2: taking the neighborhood of each feature point in each SAR image as each pixel block, and constructing the similarity between any two SAR images based on the gray value difference between the matched pixel blocks in any two SAR images.

[0049] Although there are certain differences in the acquired SAR images due to the influence of satellite position, shooting time and atmospheric interference, for SAR images analyzing ground road deformation, images acquired at different times can be used as identification features based on the ground roads, verify each other, and then identify noise pixels with no distribution pattern, thereby increasing the denoising effect of the final image and avoiding the problem of image detail loss after denoising the edge areas and other areas.

[0050] (1) Perform corner detection on each SAR image using the Harris corner detection algorithm to obtain corner points in each SAR image, and record each corner point as a feature point. The Harris corner detection algorithm is a well-known technology, and the specific process will not be described in detail.

[0051] It should be noted that for corner detection of SAR images, this application only provides one corner detection method. There are many existing corner detection methods, and implementers can also use other corner detection algorithms to perform corner detection of SAR images. This application does not make any specific restrictions.

[0052] (2) Two images acquired at different satellite angles and at different times can be considered as the pixels in one image moving to obtain the other image. Pixel block matching is performed on any two SAR images, and the similarity between the two SAR images is calculated based on the matched pixel blocks:

[0053] The center of each feature point in each SAR image is The neighborhood is used as each pixel block, and the pixel blocks of any two SAR images are matched by the three-step method (TSS) to obtain the matching pixel blocks of each pixel block in one SAR image in the other SAR image. The three-step method (TSS) is a well-known technology, and the specific process is not repeated here.

[0054] It should be noted that for the matching of pixel blocks, this application only provides a block matching algorithm. There are many existing block matching algorithms, and implementers may also use other block matching algorithms to match pixel blocks. This application does not make specific restrictions. For the size of each pixel block, in other embodiments of this application, implementers can set the size of each pixel block according to actual conditions.

[0055] Furthermore, for the two matched pixel blocks, in order to highlight the features of the feature points, the grayscale value of the central pixel point of the pixel block, that is, the feature point, is used as the grayscale value of the pixel block, and the grayscale value difference of the two matched pixel blocks is calculated and recorded as the first difference; the mean of the first difference between all the matched pixel blocks in any two SAR images is calculated, and the reciprocal of the normalized value of the mean is recorded as the similarity between the any two SAR images. The smaller the first difference between the two SAR images, the higher the similarity between the two SAR images.

[0056] Step S3, obtaining each pair of matched edge pixel points between the edge images of the arbitrary two SAR images and the similarity between the matched edge pixel points through a matching algorithm; based on the similarity and the matching situation of each edge pixel point in each edge image, combined with the similarity between the arbitrary two SAR images, calculating the first possibility that each edge pixel point is a noise pixel point.

[0057] (1) Perform edge detection on each SAR image using the Canny operator to obtain an edge image of each SAR image. The Canny operator is a well-known technology and the specific process will not be described in detail.

[0058] It should be noted that for edge detection of SAR images, this application only provides one edge detection method. There are many existing edge detection methods, and implementers can also use other edge detection algorithms to perform edge detection on SAR images. This application does not make specific restrictions.

[0059] (2) Matching analysis is also performed on the edge pixels in the edge image, so as to calculate the possibility that the edge pixels obtained in the image before denoising are noise pixels. Since edge line offset and other situations may occur in different SAR images, for any two edge images, any one of the edge images is used as a template, and the other edge image and the template are used as inputs of the shape context algorithm, and the output is the shape context between the two edge images; the shape context between the two edge images is used as the input of the Hungarian algorithm to match the edge pixels, and the output is each pair of matched edge pixels between the two edge images and the similarity between the matched edge pixels. The edge pixel matching by combining the shape context with the Hungarian algorithm is a well-known technology, and the specific process will not be repeated.

[0060] It should be noted that for the matching of edge pixels in any two edge images, this application only provides an edge pixel matching method. There are many existing edge pixel matching methods, and implementers can also use other matching algorithms for edge pixel matching. This application does not make any specific restrictions.

[0061] (3) If the current edge pixel in the target edge image does not find a matching edge pixel when the Hungarian algorithm is used to match the edge pixel in a certain edge image, the certain edge image is regarded as an unmatched image for the current edge pixel.

[0062] (4) The first probability expression of each edge pixel being a noise pixel is:

[0063]

[0064]

[0065] In the formula, Indicates the first possibility that the u-th edge pixel in the a-th edge image is a noise pixel, is the normalization function, represents the number of unmatched images of the u-th edge pixel in the a-th edge image, S represents the number of collected SAR images, represents the weighted normalized similarity value between the a-th SAR image and the b-th SAR image, represents the similarity between the u-th edge pixel in the a-th edge image and the matching edge pixel in the b-th edge image, is the similarity between the a-th SAR image and the b-th SAR image, where the edge image of the b-th SAR image is the matchable image of the u-th edge pixel in the edge image of the a-th SAR image.

[0066] The more unmatched images there are for the current edge pixel, and the smaller the similarity between the current edge pixel and the corresponding matching pixel in the image where the matching pixel can be found, the greater the first possibility that the current edge pixel is a noise pixel.

[0067] Step S4, determining a second possibility that each edge pixel is a noise pixel based on the length of the edge line where each edge pixel in the edge image is located and the number of edge pixels in the neighborhood, and calculating the noise level of each edge pixel in combination with the first possibility.

[0068] (1) Further analysis and processing is performed based on the features of the image edge. The most important road change is the change of the road line, which plays a vital role in analyzing road deformation. Since most roads are straight lines, Hough line detection is performed on each edge image to obtain the straight lines in each edge image. Among them, Hough line detection is a well-known technology and the specific process will not be repeated.

[0069] It should be noted that for straight line detection in edge images, this application only provides a straight line detection method. There are many existing straight line detection methods, and implementers can also use other straight line detection algorithms to perform straight line detection in edge images. This application does not make specific restrictions.

[0070] A length threshold is set. Preferably, in one embodiment of the present application, the length threshold is set to 30 pixel units. In other embodiments of the present application, the implementer can set the length threshold according to the actual situation. The edge pixels on the straight line in the edge image whose length is greater than the length threshold are recorded as road edge pixels. Since the distribution of noise points is random, it is almost impossible for the noise points to be connected into a line that is too long. Therefore, these road edge pixels are edge pixels where no sudden change in pixel grayscale value occurs before and after filtering. Therefore, the second possibility that these road edge pixels are noise pixels is set to 0.

[0071] (2) Further analysis is performed on edge pixels other than road edge pixels to analyze the discreteness of these edge pixels. The random appearance of noise pixels reduces the possibility that noise pixels in the image will randomly appear around edge pixels. That is, the more edge pixels there are around an edge pixel, the greater the possibility that it is a real edge pixel. If it is an edge pixel formed by noise, the number of edge pixels around it will be smaller.

[0072] Therefore, in each edge image, for other edge pixels except the road edge pixels, a neighborhood of the target edge pixel is constructed with the target edge pixel as the center. Preferably, in one embodiment of the present application, the neighborhood size is set to In other embodiments of the present application, the implementer may set the neighborhood size according to actual conditions. The number of edge pixels in the neighborhood of the target edge pixel that are not collinear with the target edge pixel is counted, which is recorded as the first number; at the same time, the length of the edge line where the target edge pixel is located is obtained, which is recorded as the first length.

[0073] The shorter the edge line where the target edge pixel is located and the fewer other edge pixels exist in the neighborhood, the greater the possibility that there are noise pixels in the area where the detected target edge pixel is located, and the greater the possibility that the corresponding target pixel is not an edge pixel.

[0074] For edge pixels other than road edge pixels, the second possibility that each edge pixel is a noise pixel is determined based on the first number and the first length of each pixel, wherein the second possibility of each edge pixel is negatively correlated with the first number and the first length of each edge pixel, respectively.

[0075] It can be understood that the negative correlation described in this application is that the independent variable decreases as the dependent variable increases, and the independent variable increases as the dependent variable decreases. The specific negative correlation can be determined according to actual conditions during application, and this application does not impose any special restrictions.

[0076] Preferably, in one embodiment of the present application, the expression of the second possibility of each edge pixel point can be: , where Im represents the second possibility of the target edge pixel, E represents the first number of the target edge pixel, and L represents the first length of the target edge pixel. is an exponential function with the natural number e as base.

[0077] In other embodiments of the present application, the second possibility expression of the target edge pixel point may be: , where is the normalization function.

[0078] (3) The average of the first possibility and the second possibility of each edge pixel is taken as the noise level of each edge pixel.

[0079] Step S5, calculating the final grayscale value of each edge pixel based on the noise level of each edge pixel and the grayscale value of the corresponding pixel in the SAR image before and after denoising, and enhancing the denoised SAR image based on the final grayscale value.

[0080] (1) Using the Lee algorithm to denoise each acquired SAR image to obtain each denoised SAR image. Preferably, in the embodiment of the present application, the filter window size of the Lee algorithm is set to As other embodiments of the present application, the implementer can set the filter window size of the Lee algorithm according to the actual situation. The Lee algorithm is a well-known technology, and the specific process will not be repeated.

[0081] It should be noted that for SAR image denoising, this application only provides one denoising method. There are many existing denoising methods, and implementers can also use other denoising algorithms to denoise SAR images. This application does not make specific restrictions.

[0082] (2) According to the noise level of each edge pixel, combined with the grayscale value before and after denoising of the corresponding pixel in the SAR image, the final grayscale value of each edge pixel is calculated. The expression is:

[0083] , where Represents the final grayscale value of the current edge pixel. Indicates the grayscale value of the pixel in the SAR image before denoising corresponding to the current edge pixel. Indicates the gray value of the pixel in the denoised SAR image corresponding to the current edge pixel. Indicates the noise level of the current edge pixel.

[0084] (3) The final grayscale value of each edge pixel is used to replace the grayscale value of the corresponding pixel in the denoised SAR image, thereby enhancing the denoised SAR image and improving the ability of the Lee algorithm to retain edge information.

[0085] When correcting the grayscale value of edge pixels, the lower the noise content of the edge pixels, the closer the grayscale value after correction is to the grayscale value of the pixel before denoising. If the noise content is higher, the grayscale value after correction is closer to the grayscale value after using the Lee algorithm.

[0086] Step S6: Based on the enhanced denoised images of all SAR images, urban road deformation monitoring is performed in combination with a surface deformation detection algorithm.

[0087] The enhanced denoised images of all SAR images are used as inputs of the PS-InSAR algorithm, and the outputs are the deformation variables of each point on the urban road. The PS-InSAR algorithm is a well-known technology, and the specific process will not be described in detail.

[0088] It should be noted that for the deformation detection of each point on the urban road, this application only provides a surface deformation detection method. There are many existing surface deformation detection methods. Implementers can also use other surface deformation detection algorithms to detect the deformation of each point on the urban road. This application does not make specific restrictions.

[0089] A deformation threshold is set. Preferably, in one embodiment of the present application, the deformation threshold is set to 9 mm. In other embodiments of the present application, the implementer can set the deformation threshold according to the actual situation. If the deformation amount of any point in the urban road is greater than the deformation threshold, an alarm is performed; if the deformation amount of each point in the urban road is less than or equal to the deformation threshold, no alarm is performed.

[0090] See also Figure 2 , Figure 2 1 is a schematic diagram of the structure of a deformation monitoring system for urban road engineering provided in an embodiment of the present application. In this embodiment, each unit included in the terminal is used to execute each step in an embodiment corresponding to a deformation monitoring method for urban road engineering. Figure 2 , the urban road deformation monitoring system includes:

[0091] Data acquisition module: collects various SAR images of urban roads;

[0092] Image similarity analysis module: taking the neighborhood of each feature point in each SAR image as each pixel block, and constructing the similarity between any two SAR images based on the gray value difference between the matching pixel blocks in any two SAR images;

[0093] Noise possibility assessment module: obtains each pair of matched edge pixel points between the edge images of any two SAR images and the similarity between the matched edge pixel points through a matching algorithm; calculates the first possibility that each edge pixel point is a noise pixel point based on the similarity and the matching situation of each edge pixel point in each edge image and the similarity between the any two SAR images;

[0094] Noise degree calculation module: determines the second possibility that each edge pixel is a noise pixel based on the length of the edge line where each edge pixel is located in the edge image and the number of edge pixels in the neighborhood, and calculates the noise degree of each edge pixel in combination with the first possibility;

[0095] Image enhancement module: calculates the final grayscale value of each edge pixel based on the noise level of each edge pixel and the grayscale value of the corresponding pixel in the SAR image before and after denoising, and enhances the denoised SAR image based on the final grayscale value;

[0096] Urban road deformation monitoring module: Based on the enhanced denoised images of all SAR images, urban road deformation monitoring is performed in combination with the surface deformation detection algorithm.

[0097] Based on the same inventive concept as the above method, an embodiment of the present application also provides a deformation monitoring device for an urban road project, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the computer program, the steps of any one of the above-mentioned methods for deformation monitoring of an urban road project are implemented.

[0098] In summary, the embodiment of the present application provides a deformation monitoring method for an urban road project, which calculates the similarity between any two urban road SAR images by the similarity between the feature points in any two urban road SAR images; for the edge images of all urban road SAR images, the first possibility that each edge pixel is a noise pixel is calculated based on the matching degree of edge pixels in any two edge images and the similarity between the corresponding two urban road SAR images; the second possibility that each edge pixel is a noise pixel is calculated based on the irregular distribution characteristics of the noise pixels; the grayscale value of the edge pixel is corrected based on the above characteristics and the grayscale value of the edge pixel before and after denoising, and the image filtered by the Lee algorithm is enhanced based on the corrected grayscale value, so that the filtered image can maintain its denoising effect and greatly retain its edge information, which greatly increases the denoising effect of the Lee algorithm; based on the enhanced denoised SAR image of the urban road, the urban road deformation monitoring is performed in combination with the surface deformation detection algorithm, thereby improving the accuracy and reliability of the urban road deformation monitoring results.

[0099] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. The above-mentioned specific embodiments of the present application are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0100] The various embodiments in the present application are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0101] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present application should be included in the protection scope of the present application.

Claims

1. A deformation monitoring method for urban road engineering, characterized in that: The method comprises the following steps: Collect various SAR images of urban roads; Taking the neighborhood of each feature point in each SAR image as each pixel block, and constructing the similarity between any two SAR images based on the gray value difference between the matched pixel blocks in any two SAR images; Obtaining each pair of matched edge pixel points between the edge images of any two SAR images and the similarity between the matched edge pixel points through a matching algorithm; calculating a first possibility that each edge pixel point is a noise pixel point based on the similarity between the edge pixel points and the matching condition of each edge pixel point in each edge image and in combination with the similarity between the any two SAR images; Determine a second possibility that each edge pixel is a noise pixel based on the length of the edge line where each edge pixel is located and the number of edge pixels in the neighborhood in the edge image, and calculate the noise level of each edge pixel in combination with the first possibility; Calculating a final grayscale value of each edge pixel based on the noise level of each edge pixel and the grayscale values ​​of the corresponding pixel in the SAR image before and after denoising, and enhancing the denoised SAR image based on the final grayscale value; Based on the enhanced denoised images of all SAR images, urban road deformation monitoring is carried out in combination with the surface deformation detection algorithm.

2. A deformation monitoring method for urban road engineering as claimed in claim 1, characterized in that: The process of obtaining the similarity between any two SAR images is as follows: Matching the pixel blocks of any two SAR images by a block matching algorithm to obtain matching pixel blocks of each pixel block in one SAR image in the other SAR image; The gray value difference between the central pixel points of the two matched pixel blocks is recorded as a first difference; the mean of all the first differences between the arbitrary two SAR images is calculated, and the reciprocal of the normalized value of the mean is recorded as the similarity between the arbitrary two SAR images.

3. A deformation monitoring method for urban road engineering as claimed in claim 1, characterized in that: The process of obtaining the first possibility that each edge pixel is a noise pixel is as follows: If the current edge pixel does not have a matching edge pixel in a certain edge image, the certain edge image is used as the unmatched image of the current edge pixel; the first possibility that the uth edge pixel in the ath edge image is a noise pixel is recorded as , The expression is: In the formula, is the normalization function, represents the number of unmatched images of the u-th edge pixel in the a-th edge image, S represents the number of collected SAR images, represents the normalized value of the similarity between the a-th and b-th SAR images, Represents the similarity between the u-th edge pixel in the a-th edge image and the matching edge pixel in the b-th edge image.

4. A deformation monitoring method for urban road engineering as claimed in claim 1, characterized in that: The process of obtaining the second possibility that each edge pixel is a noise pixel is as follows: The straight lines in each edge image are obtained by a straight line detection algorithm, and the edge pixel points on the straight line whose length is greater than a preset length threshold are recorded as road edge pixel points; the second possibility that the road edge pixel point is a noise pixel point is set to 0; For edge pixels other than road edge pixels, the number of edge pixels that are not collinear with the target edge pixel in the neighborhood of the target edge pixel is counted, recorded as the first number, the length of the edge line where the target edge pixel is located is obtained, recorded as the first length, and the second possibility of each edge pixel being a noise pixel is determined based on the first number and the first length, wherein the second possibility of each edge pixel is negatively correlated with the first number and the first length of each edge pixel, respectively.

5. A deformation monitoring method for urban road engineering as claimed in claim 1, characterized in that: The noise level of each edge pixel point is: the average of the first possibility and the second possibility of each edge pixel point.

6. A deformation monitoring method for urban road engineering as claimed in claim 1, characterized in that: The final grayscale value of each edge pixel is expressed as: , where Represents the final grayscale value of the current edge pixel. and Respectively represent the grayscale value of the pixel in the SAR image before and after denoising corresponding to the current edge pixel, Indicates the noise level of the current edge pixel.

7. A deformation monitoring method for urban road engineering as claimed in claim 1, characterized in that: The method for enhancing the denoised SAR image is: using the final grayscale value of each edge pixel to replace the grayscale value of the corresponding pixel in the denoised SAR image.

8. A deformation monitoring method for urban road engineering as claimed in claim 1, characterized in that: The urban road deformation monitoring is specifically as follows: The enhanced and denoised images of all SAR images are used as the input of the surface deformation detection algorithm, and the output is the deformation variable of each point on the urban road; if the deformation variable of any point in the urban road is greater than the deformation threshold, an alarm is performed; otherwise, no alarm is performed.

9. A deformation monitoring system for urban road engineering, implementing the method according to claim 1, characterized in that: The system comprises: Data acquisition module: collects various SAR images of urban roads; Image similarity analysis module: taking the neighborhood of each feature point in each SAR image as each pixel block, and constructing the similarity between any two SAR images based on the gray value difference between the matching pixel blocks in any two SAR images; Noise possibility assessment module: obtains each pair of matched edge pixel points between the edge images of any two SAR images and the similarity between the matched edge pixel points through a matching algorithm; calculates the first possibility that each edge pixel point is a noise pixel point based on the similarity between the edge pixel points and the matching conditions of each edge pixel point in each edge image, combined with the similarity between the any two SAR images; Noise degree calculation module: determines the second possibility that each edge pixel is a noise pixel based on the length of the edge line where each edge pixel is located in the edge image and the number of edge pixels in the neighborhood, and calculates the noise degree of each edge pixel in combination with the first possibility; Image enhancement module: calculates the final grayscale value of each edge pixel based on the noise level of each edge pixel and the grayscale value of the corresponding pixel in the SAR image before and after denoising, and enhances the denoised SAR image based on the final grayscale value; Urban road deformation monitoring module: Based on the enhanced denoised images of all SAR images, urban road deformation monitoring is performed in combination with the surface deformation detection algorithm.

10. A deformation monitoring device for urban road engineering, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images