A common highway slope deformation time sequence monitoring method based on TL-1 satellite

Abstract

The application relates to the technical field of image analysis, in particular to a common highway slope deformation time sequence monitoring method based on a TL-1 satellite, which comprises the following steps: constructing a three-dimensional model of a target region based on historical images and a coordinate system; inputting each satellite image in a time sequence image into an image recognition model to obtain a corresponding recognition image, extracting a crack region in the target region based on each recognition image, obtaining crack information in each satellite image, and obtaining first deformation time sequence information according to crack changes; updating each satellite image into the three-dimensional model according to a time sequence, obtaining displacement change information and morphological change information of the target region in a preset time period, obtaining second deformation time sequence information based on the displacement change information and the morphological change information; and obtaining a slope deformation prediction result according to the first deformation time sequence information, the second deformation time sequence information and a deformation prediction model. The application solves the problem of poor slope deformation recognition precision and improves the monitoring precision of slope deformation.

Description

Technical Field

[0001] This application relates to the field of image analysis technology, and in particular to a time-series monitoring method for deformation of ordinary highway slopes based on the TL-1 satellite. Background Technology

[0002] Ordinary highway slopes are at risk of deformation and landslides due to factors such as natural erosion, climate change, and human activities. Traditional slope monitoring methods rely on ground sensors and manual inspections, which suffer from high costs, untimely data collection, and limited coverage. With the rapid development of transportation infrastructure, the stability problem of ordinary highway slopes is becoming increasingly prominent, and traditional slope monitoring methods suffer from high costs, low efficiency, and poor real-time performance. Therefore, it is urgent to develop an efficient and real-time slope deformation monitoring method. A similar prior art is Chinese patent application CN118552848A, which proposes a method, device, electronic device, and storage medium for monitoring slope deformation. The method includes: acquiring monitoring images of the slope monitoring area using an image acquisition device; determining the current illumination type of the slope monitoring area based on the monitoring images; determining a template image of the slope monitoring area under the illumination type based on the acquisition time of the monitoring images and the illumination type; and comparing the monitoring images with the template image to determine whether deformation exists in the slope monitoring area. The method and apparatus described above avoid shadow variations caused by differences in light direction and intensity, reduce false alarms due to light changes, and improve the accuracy of slope deformation monitoring. Furthermore, a similar prior art exists in Japanese patent application JP2023140677A, which provides a slope instability location extraction system. This system includes: an acquisition unit for acquiring a visible light image of the collapse extraction area on the ground surface and a vegetation index map representing the vegetation index distribution in the collapse extraction area; a low-index detection process; areas in the vegetation index map with vegetation indices less than a threshold; and an extraction process for extracting slope collapse locations by performing pattern matching on the target areas in the visible light image that include low-index areas. This system can extract slope instability locations at low cost while improving extraction accuracy. Although the two patent applications mentioned above have solved the problem of slope monitoring and identification, they do not obtain deformation characteristics within a preset time period based on the acquisition time sequence, nor do they obtain specific deformation categories and corresponding deformation trends based on the deformation characteristics. Therefore, they cannot obtain the development trend of slope deformation. The purpose of this invention is to provide a time-series monitoring method for ordinary highway slope deformation based on LT-1 satellite. This method can achieve high-precision prediction of slope deformation and improve the reliability and safety of monitoring. Summary of the Invention

[0003] To better address the aforementioned problems, this invention provides a time-series monitoring method for deformation of ordinary highway slopes based on the LT-1 satellite. The method includes:

[0004] Historical images of the target area are acquired, the historical images are preprocessed, a fixed object outside the target area is selected as a reference object, and the reference object is used as the origin of the coordinate system. A three-dimensional model of the target area is constructed based on the historical images and the origin of the coordinate system.

[0005] Each satellite image in the time-series images acquired within a preset time period is input into the image recognition model to obtain the recognition image corresponding to the satellite image. Based on each recognition image, the crack region in the target region is extracted, and the crack information corresponding to the crack region in each satellite image is obtained. The crack change is obtained through all the crack information corresponding to the time-series images, and the first deformation time-series information is obtained based on the crack change.

[0006] Each satellite image in the time series is updated to the three-dimensional model in chronological order, and the position and shape information of the slope in the target area in each satellite image are obtained. Based on the position and shape information of the slope corresponding to each satellite image, the displacement and shape change information of the target area within the preset time period are obtained. Based on the displacement and shape change information, the second deformation time series information is obtained.

[0007] The slope deformation prediction results are obtained based on the first deformation time series information, the second deformation time series information, and the deformation prediction model.

[0008] As a preferred embodiment of the present invention, the acquisition of crack information includes:

[0009] Each satellite image in the time-series images acquired within the preset time period is preprocessed and input into the image recognition model to obtain the recognition image corresponding to each satellite image; wherein, the recognition image is a binary image, and is an image that distinguishes the crack area and the slope area in the corresponding target area by pixel value, and the satellite image is a three-dimensional image;

[0010] Crack regions are extracted based on the crack probability values ​​corresponding to cracks in the identified image. Each crack region is divided into multiple sub-segments according to a set vertical distance. The first and second tangents on both sides of each sub-segment are obtained. A first and second angle are calculated between each adjacent sub-segment and the first and second tangents on the same side of the sub-segment. The first and second angles are angles within the crack region. If at least one of the first and second angles between at least one adjacent sub-segment and the sub-segment is less than a set angle, the sub-segment is considered a crack segment. When both the first included angle and the second included angle are greater than or equal to a set angle, at least one of the adjacent sub-segments is taken as the target sub-segment, and the target sub-segment and the sub-segment are merged into a new sub-segment. The two sides of the first included angle are connected to the two intersection points of the corresponding sub-segment and used as the third tangent of the new sub-segment. Similarly, the fourth tangent on the other side of the new sub-segment is obtained, and this step is repeated to obtain multiple crack segments of the crack region. The crack included angle between adjacent crack segments is greater than or equal to the set angle. The crack region in the recognition image, the crack segment corresponding to each crack region, and the size and position information of each crack segment are used as the crack information.

[0011] As a preferred embodiment of the present invention, the acquisition of the first deformation timing information of the target region includes:

[0012] Crack change information is obtained by comparing the crack segment in the crack information corresponding to each satellite image in the time-series image with the crack information of the crack segment at the same position in the crack information corresponding to the adjacent satellite images in the time sequence, wherein the acquisition time of the adjacent satellite images is later than the acquisition time of the satellite images. Based on the crack change information corresponding to each satellite image and the adjacent satellite images, the first deformation time-series information of the target area in the time-series image within the preset time period is obtained, wherein the crack change information includes size change and depth change.

[0013] As a preferred embodiment of the present invention, the training of the image recognition model includes:

[0014] The image recognition model is created by the model creation unit and a sample image is read from the storage unit. The sample image includes first label information, which is a pre-labeled label. The sample image and the first label information are used as first training data to train the image recognition model. During the training iteration, the average moving weight of the prediction result corresponding to each pixel in the sample image is periodically calculated, and the average moving weight is used to replace the prediction result obtained in the current iteration. Second label information is set for second pixels whose prediction result is greater than a set threshold. The second pixel is a pixel that is not labeled with the pre-labeled label, and the pixel that includes the pre-labeled label is used as the first pixel.

[0015] At least two second pixels adjacent to the first pixels are obtained, and at least two first pixels are used as reference pixels. The first label information corresponding to the reference pixels is the same and located on both sides of the second pixel. The second label information corresponding to the second pixel is compared with the first label information corresponding to the reference pixel. When the second label information is the same as the first label information corresponding to the reference pixel, the second pixel and the corresponding second label information are added as first learning sample pixels to the second training data. Otherwise, the difference features between the second pixel and the reference pixels are extracted again and enhanced. The image recognition model is then retrained using the second training data.

[0016] As a preferred embodiment of the present invention, the acquisition of the second deformation timing information includes:

[0017] Each preprocessed satellite image acquired within the preset time period is sequentially updated to the three-dimensional model. The position and shape information of the slope in the temporally adjacent satellite images are compared to obtain displacement change information and slope shape information change, and the displacement change information and slope shape change information are used as the second deformation temporal information.

[0018] As a preferred technical solution of the present invention, obtaining the slope deformation prediction result includes:

[0019] The first deformation time series information and the second deformation time series information are fused according to time series to obtain slope deformation time series information, and the slope deformation time series information is input into the deformation prediction model to obtain the slope deformation prediction result. The slope deformation prediction result includes the future deformation development trend and slope deformation type of the target area.

[0020] As a preferred technical solution of the present invention, the deformation prediction model is a model trained with historical deformation data. The historical deformation data includes displacement deformation, crack deformation and corresponding deformation types, as well as the entire development process of displacement deformation and crack deformation during the entire deformation process of the deformation type.

[0021] As a preferred technical solution of the present invention, the deformation development trend of the target area includes the displacement change trend, morphological change trend and crack change trend of the slope over time.

[0022] This invention also provides a time-series monitoring system for deformation of ordinary highway slopes based on TL-1, the system being used to implement the above-described method, the system comprising:

[0023] The model creation unit is used to acquire historical images of the target area, preprocess the historical images, select a fixed object outside the target area as a reference object, and use the reference object as the origin of the coordinate system to construct a three-dimensional model of the target area based on the historical images and the origin of the coordinate system.

[0024] The computing unit is configured to: input each satellite image in the time-series images acquired within the preset time period into an image recognition model to obtain a recognition image corresponding to the satellite image; extract the crack region in the target region based on each recognition image and obtain crack information corresponding to the crack region in each satellite image; obtain crack changes through all the crack information corresponding to the time-series images and obtain first deformation time-series information based on the crack changes; update each satellite image in the time-series images to the three-dimensional model in chronological order and obtain the position and shape information of the slope in the target region in each satellite image; obtain the displacement and shape change information of the target region within the preset time period based on the position and shape information of the slope corresponding to each satellite image; and obtain second deformation time-series information based on the displacement and shape change information.

[0025] The prediction unit is used to obtain the slope deformation prediction result based on the first deformation time series information, the second deformation time series information, and the deformation prediction model.

[0026] The present invention provides a computer storage medium, characterized in that the storage medium stores program instructions, wherein the program instructions, when executed, control the device where the storage medium is located to perform the above-described method.

[0027] Compared with the prior art, the beneficial effects of the present invention are at least as follows:

[0028] This invention constructs a three-dimensional model of the target area by combining historical and time-series images from the LT-1 satellite, and extracts information on crack areas and their changes, achieving high-precision monitoring of deformation on ordinary highway slopes. This method provides more accurate deformation data, improving the accuracy and reliability of monitoring compared to traditional methods. Utilizing time-series images from the LT-1 satellite, this invention can update displacement and morphological changes in the target area in real time, obtaining the latest deformation time-series information. This real-time monitoring capability enables management departments to respond promptly to slope deformation, improving the timeliness of geological disaster response. It automatically extracts crack information through image recognition models, ensuring the accuracy of these models even without the need for precise annotation of sample images. This reduces labor intensity and human error, improving the efficiency and accuracy of crack analysis. By combining first and second deformation time-series information, and through the synergy of the above technical solutions, it can not only predict the future deformation trend and type of slopes, providing a scientific basis for preventing and mitigating geological disasters, but also reduce reliance on ground sensors and manual inspections, lowering monitoring costs while improving monitoring efficiency. Especially in remote or inaccessible areas, this invention not only focuses on crack deformation but also incorporates displacement and morphological change information. Through data fusion technology, it provides a more comprehensive deformation analysis. The method of this invention is applicable to slope deformation monitoring under various environmental conditions, unrestricted by weather, terrain, or other natural conditions, and has strong environmental adaptability. Attached Figure Description

[0029] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 This is a flowchart illustrating the time-series monitoring method for ordinary highway slope deformation based on LT-1 satellite in the embodiments of this application.

[0031] Figure 2 This is a schematic diagram of the structure of the ordinary highway slope deformation time-series monitoring system based on the LT-1 satellite in the embodiments of this application. Detailed Implementation

[0032] This application provides a method, apparatus, device, and medium for customizing feed formulations based on artificial intelligence. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0033] For ease of understanding, the specific process of the embodiments of this application is described below, such as... Figure 1 As shown in the embodiments of this application, the time-series monitoring method for ordinary highway slope deformation based on LT-1 satellite includes:

[0034] Step S1: Obtain historical images of the target area, preprocess the historical images, select a fixed object outside the target area as a reference object, and use the reference object as the origin of the coordinate system to construct a three-dimensional model of the target area based on the historical images and the origin of the coordinate system.

[0035] Specifically, by preprocessing historical images of the target area, including angle adjustment and noise reduction, and selecting a fixed object outside the target area and at a certain distance from it as a reference, a coordinate system is established with the reference object as the origin. This ensures that even if the target area undergoes deformation, the position of the coordinate origin remains unchanged, enabling accurate measurement of slope deformation. A three-dimensional model is then established using the historical images and the coordinate system with the reference object as the origin. Through this technical solution, the three-dimensional model of the target area can be obtained, thus laying the foundation for accurately calculating crack deformation information, displacement changes, and morphological changes in the target area.

[0036] Step S2: Input each satellite image in the time series images acquired within the preset time period into the image recognition model to obtain the recognition image corresponding to the satellite image, extract the crack region in the target region based on each recognition image, and obtain the crack information corresponding to the crack region in each satellite image, obtain the crack change through each crack information corresponding to the time series image, and obtain the first deformation time series information based on the crack change;

[0037] Specifically, by inputting each satellite image in the aforementioned time-series images into the aforementioned image recognition model, the corresponding recognition image is obtained. Based on the recognition image, the slope area and crack area in the aforementioned satellite image are obtained, and the crack information of the aforementioned crack area is obtained. The crack information corresponding to the aforementioned satellite image is compared with the crack information corresponding to the adjacent satellite image according to the time sequence to obtain crack change information. Based on the crack change information between each of the aforementioned satellite images and the aforementioned adjacent satellite images within the aforementioned preset time period, the first deformation time sequence information of the aforementioned target area within the aforementioned preset time period is obtained. Through the above technical solution, the crack change information of the crack segment corresponding to each crack area in the aforementioned target area can be accurately obtained, laying the foundation for further combining the second deformation time sequence information, namely the displacement change information and morphological change information of the slope, to obtain accurate slope deformation time sequence information.

[0038] Step S3: Update each satellite image in the time series image to the three-dimensional model in chronological order, and obtain the position information and morphological information of the slope in the target area in each satellite image. Obtain the displacement change information and morphological change information of the target area within the preset time period based on the position information and morphological change information of the slope corresponding to each satellite image. Obtain the second deformation time series information based on the displacement change information and the morphological change information.

[0039] Specifically, by updating each of the satellite images obtained within the preset time period after preprocessing the time-series images to the three-dimensional model, and sequentially recording the corresponding slope position and shape in the three-dimensional model, the changes in slope position and shape in the three-dimensional model corresponding to each pair of adjacent satellite images are calculated. Based on the changes in position and shape, the displacement and shape changes of the slope within the preset time period are obtained, and the displacement and shape changes of the slope are used as the second deformation time-series information of the target area. Through the above technical solution, the displacement and shape changes of the slope in the target area within the preset time period can be accurately obtained, laying the foundation for further obtaining the slope deformation of the target area.

[0040] Step S4: Obtain the slope deformation prediction result based on the first deformation time series information, the second deformation time series information, and the deformation prediction model.

[0041] Specifically, by fusing the first deformation time-series information and the second deformation time-series information in a temporal sequence, complete time-varying slope deformation time-series information can be obtained. Based on the slope deformation time-series information and the deformation prediction model, the future deformation development trend and edge deformation type of the target area can be obtained. Through the above technical solution, the deformation development trend and deformation type of the target area can be understood in a timely manner, providing a basis for taking timely intervention measures to prevent deformation deterioration. Further, the acquisition of crack information includes:

[0042] Each satellite image in the time-series images acquired within the preset time period is preprocessed and input into the image recognition model to obtain the recognition image corresponding to each satellite image; wherein, the recognition image is a binary image, and is an image that distinguishes the crack area and the slope area in the corresponding target area by pixel value, and the satellite image is a three-dimensional image;

[0043] Specifically, the aforementioned preset time period is the time period during which the LT-1 satellite can capture images of the aforementioned target area. The aforementioned time-series image is all satellite images captured in chronological order at a set period within the aforementioned preset time period. Each of the aforementioned satellite images in the aforementioned time-series image is preprocessed and input into the aforementioned image recognition model. The aforementioned preprocessing includes enhancement, noise reduction, angle adjustment, and light correction. Through the aforementioned angle correction, a frontal image of the aforementioned target area can be obtained, i.e., the slope direction where the crack is located, thereby facilitating crack identification. Through the aforementioned image recognition model, the probability of each pixel being classified as a slope or a crack can be obtained, i.e., the aforementioned recognition image. Since the aforementioned recognition image is a binary image, the slope and crack can be distinguished more clearly. Through the aforementioned technical solution, an accurate recognition image can be obtained, laying the foundation for further extraction of crack information.

[0044] Crack regions are extracted based on the crack probability values ​​corresponding to cracks in the identified image. Each crack region is divided into multiple sub-segments according to a set vertical distance. The first and second tangents on both sides of each sub-segment are obtained. A first and second angle are calculated between each adjacent sub-segment and the first and second tangents on the same side of the sub-segment. The first and second angles are angles within the crack region. If at least one of the first and second angles between at least one adjacent sub-segment and the sub-segment is less than a set angle, the sub-segment is considered a crack segment. When both the first included angle and the second included angle are greater than or equal to a set angle, at least one of the adjacent sub-segments is taken as the target sub-segment, and the target sub-segment and the sub-segment are merged into a new sub-segment. The two sides of the first included angle are connected to the two intersection points of the corresponding sub-segment and used as the third tangent of the new sub-segment. Similarly, the fourth tangent on the other side of the new sub-segment is obtained, and this step is repeated to obtain multiple crack segments of the crack region. The crack included angle between adjacent crack segments is greater than or equal to the set angle. The crack region in the recognition image, the crack segment corresponding to each crack region, and the size and position information of each crack segment are used as the crack information.

[0045] Specifically, pixel regions in the aforementioned identification image whose crack probability value is greater than a first set threshold are designated as crack regions, and pixel regions in the aforementioned identification image whose slope probability value is greater than a second set threshold are designated as slope regions. Both the first and second set thresholds are greater than or equal to 0.8, and the maximum value of the crack probability value and the slope probability value is 1. Each crack region in the aforementioned identification image is further divided into multiple sub-segments according to a set vertical distance, where the set distance is the length of k pixel rows, and k is greater than or equal to 3 and less than or equal to 5. The first and second tangents on both sides of each sub-segment are obtained. The first and second tangents are line segments on the corresponding side that are tangent to pixels closest to the other side and intersect with the corresponding side. The first angle between the first tangents on the same side of two adjacent sub-segments and the second angle between the second tangents on the same side are calculated. Both the first and second angles are angles pointing towards the crack region. In two adjacent sub-segments of the aforementioned sub-segment, the first and second angles corresponding to each adjacent sub-segment are considered to be... When at least one of the angles is less than a set angle, i.e., when the crack trends of the aforementioned sub-segment and its two adjacent sub-segments are inconsistent, the aforementioned sub-segment is designated as a crack segment. When at least one of the adjacent sub-segments of the aforementioned sub-segment has an angle greater than the set angle, at least one adjacent sub-segment is merged with the aforementioned sub-segment into a new sub-segment. The two sides of the aforementioned first angle are connected to the intersection points of the corresponding two sub-segments to obtain the third tangent of the new sub-segment. Similarly, the fourth tangent of the second angle corresponding to the new sub-segment is obtained. The method of obtaining the crack segments is repeated to obtain the crack segments in each of the aforementioned crack regions. The scale information and location information of each of the aforementioned crack regions in the aforementioned identification image, the crack segments corresponding to each crack region, and each of the aforementioned crack segments are used as the aforementioned crack information. Through the above technical solution, not only can the identification image of each satellite image in the aforementioned time series image be obtained, but also the crack regions and crack segments of each crack region in the aforementioned identification image can be obtained, thereby laying the foundation for further obtaining the change trend of the cracks corresponding to the aforementioned time series image and obtaining the aforementioned first deformation time series information.

[0046] Further, the acquisition of the first deformation timing information of the target region includes:

[0047] Crack change information is obtained by comparing the crack segment in the crack information corresponding to each satellite image in the time-series image with the crack information of the crack segment at the same position in the crack information corresponding to the adjacent satellite images in the time sequence, wherein the acquisition time of the adjacent satellite images is later than the acquisition time of the satellite images. Based on the crack change information corresponding to each satellite image and the adjacent satellite images, the first deformation time-series information of the target area in the time-series image within the preset time period is obtained, wherein the crack change information includes size change and depth change.

[0048] Specifically, by comparing the crack information corresponding to each satellite image in the aforementioned time-series image with the crack information corresponding to adjacent satellite images according to the acquisition time sequence, crack change information is obtained. Based on the crack change information between each of the aforementioned satellite images and the aforementioned adjacent satellite images within the aforementioned preset time period, the aforementioned first deformation time-series information of the aforementioned target area within the aforementioned preset time period is obtained. Through the above technical solution, the crack change information of the crack segment corresponding to each crack area in the aforementioned target area can be accurately obtained, laying the foundation for further combining the second deformation time-series information, namely the displacement change of the slope, to obtain accurate slope deformation time-series information.

[0049] Furthermore, the training of the image recognition model includes:

[0050] The image recognition model is created by the model creation unit and a sample image is read from the storage unit. The sample image includes first label information, which is a pre-labeled label. The sample image and the first label information are used as first training data to train the image recognition model. During the training iteration, the average moving weight of the prediction result corresponding to each pixel in the sample image is periodically calculated, and the average moving weight is used to replace the prediction result obtained in the current iteration. Second label information is set for second pixels whose prediction result is greater than a set threshold. The second pixel is a pixel that is not labeled with the pre-labeled label, and the pixel that includes the pre-labeled label is used as the first pixel.

[0051] Specifically, since pixel-level annotation of the sample images would be labor-intensive and require significant manpower, a pre-annotation method is adopted, i.e., partial annotation. Pre-labels, namely the first label information, are added to the sample images. This first label information includes slope pixel labels corresponding to some slope areas and crack pixel labels corresponding to some crack areas in the sample images. The sample images and their corresponding first label information are used as the first training data for the image recognition model. During training, the average moving weight of the prediction result for each pixel in the sample images is periodically calculated; that is, the weighted sum of the current prediction value and the previous average moving weight, where the weight of the current prediction result is greater than that of the previous prediction. The moving average weight is used to replace the prediction result, and the second label information is set for the second pixel whose prediction result is greater than a set threshold. The prediction result is the maximum value between the first probability that the pixel is a slope pixel and the second probability that it is a crack pixel. For example, if the probability that the second pixel is a slope is 0.85, then the second label information is "the second pixel is a slope pixel". The second pixel is a pixel that is not labeled with the pre-labeled label, that is, a pixel of unsupervised learning. Through the above technical solution, the predicted label of the unlabeled pixel, that is, the second label information, is obtained, which lays the foundation for further learning based on the second label information and the corresponding second pixel.

[0052] At least two second pixels adjacent to the first pixels are obtained, and at least two first pixels are used as reference pixels. The first label information corresponding to the reference pixels is the same and located on both sides of the second pixel. The second label information corresponding to the second pixel is compared with the first label information corresponding to the reference pixel. When the second label information is the same as the first label information corresponding to the reference pixel, the second pixel and the corresponding second label information are added as first learning sample pixels to the second training data. Otherwise, the difference features between the second pixel and the reference pixels are extracted again and enhanced. The image recognition model is then retrained using the second training data.

[0053] Specifically, since the first pixel is trained under supervised conditions, its prediction result is more accurate than that of the second pixel. However, because the second pixel adjacent to a single first pixel may be located at the boundary between a slope and a crack, it's impossible to accurately determine whether the second label information of the second pixel is the same as the first label information of the first pixel. Therefore, by acquiring the second pixels adjacent to at least two of the first pixels and using the first pixels as reference pixels, the first label information corresponding to the reference pixels is compared with the second label information. The reference pixels are located on either side of the second pixels. When the second label information of the second pixel is the same as the first label information of the reference pixel, it indicates that the second pixel... If the predicted result of the pixel, i.e., the second label information, is accurate, then the second pixel and its corresponding second label information are used as the second training data. Conversely, if the second label information of the second pixel is different from the first label information of the reference pixel, it indicates that the label information of the second pixel is not accurate enough. The difference features between the second pixel and the reference pixel are then extracted and enhanced to improve the recognition between the second pixel and the reference pixel. At the same time, the image recognition model is retrained using the second training data. Through the above technical solution, the pre-labeled satellite image can be accurately identified at the pixel level, thereby accurately identifying slope areas and crack areas, thus improving the accuracy of slope deformation recognition.

[0054] Furthermore, the acquisition of the second deformation timing information includes:

[0055] Each preprocessed satellite image acquired within the preset time period is sequentially updated to the three-dimensional model. The position and shape information of the slope in the temporally adjacent satellite images are compared to obtain displacement change information and slope shape information change, and the displacement change information and slope shape change information are used as the second deformation temporal information.

[0056] Specifically, by updating each of the satellite images obtained within the preset time period after preprocessing the time-series images to the three-dimensional model, and sequentially recording the corresponding slope position and morphology information in the three-dimensional model, the slope position and morphology changes in the three-dimensional model corresponding to each pair of adjacent satellite images are calculated. Based on the position and morphology changes, the displacement and morphology changes of the slope within the preset time period are obtained, and the displacement and morphology changes of the slope are used as the second deformation time-series information of the target area. Through the above technical solution, the displacement and morphology changes of the slope in the target area within the preset time period can be accurately obtained, laying the foundation for further obtaining the slope deformation of the target area.

[0057] Furthermore, obtaining the slope deformation prediction results includes:

[0058] The first deformation time series information and the second deformation time series information are fused according to time series to obtain slope deformation time series information, and the slope deformation time series information is input into the deformation prediction model to obtain the slope deformation prediction result. The slope deformation prediction result includes the future deformation development trend and slope deformation type of the target area.

[0059] Specifically, by fusing the first deformation time series information and the second deformation time series information in a time sequence, complete time-varying slope deformation time series information can be obtained. Based on the slope deformation time series information and the deformation prediction model, the future deformation development trend and the type of edge deformation of the target area can be obtained. Through the above technical solution, the deformation development trend and deformation type of the target area can be understood in a timely manner, providing a basis for taking timely intervention measures to prevent deformation deterioration.

[0060] Furthermore, the deformation prediction model is a model trained on historical deformation data, which includes displacement deformation, slope morphology changes, crack deformation and corresponding deformation types, as well as the entire development process of displacement deformation, slope morphology changes and crack deformation during the entire deformation process of the deformation type.

[0061] Specifically, the aforementioned displacement deformation, slope morphology changes, and crack deformation are deformation information of the corresponding deformation types over time. Through the above technical solution, a high-precision deformation prediction model can be obtained, thereby providing a basis for obtaining the corresponding deformation types and subsequent slope deformation trends through the displacement and crack changes in the target area.

[0062] Furthermore, the deformation development trend of the target area includes the displacement change trend, morphological change trend, and crack change trend of the slope over time.

[0063] This invention also provides a time-series monitoring system for deformation of ordinary highway slopes based on TL-1, the system being used to implement the above-mentioned method, such as... Figure 2 As shown, the system includes:

[0064] The model creation unit is used to acquire historical images of the target area, preprocess the historical images, select a fixed object outside the target area as a reference object, and use the reference object as the origin of the coordinate system to construct a three-dimensional model of the target area based on the historical images and the origin of the coordinate system.

[0065] The computing unit is configured to: input each satellite image in the time-series images acquired within the preset time period into an image recognition model to obtain a recognition image corresponding to the satellite image; extract the crack region in the target region based on each recognition image and obtain crack information corresponding to the crack region in each satellite image; obtain the crack change trend through all the crack information corresponding to the time-series images and obtain first deformation time-series information based on the crack change trend; update each satellite image in the time-series images to the three-dimensional model in chronological order and obtain the position and shape information of the slope in the target region in each satellite image; obtain the displacement change information and shape change information of the target region within the preset time period based on the position and shape information of the slope corresponding to each satellite image; and obtain second deformation time-series information based on the displacement change information and shape change information.

[0066] The prediction unit is used to obtain the slope deformation prediction result based on the first deformation time series information, the second deformation time series information, and the deformation prediction model.

[0067] The present invention also provides a computer storage medium storing program instructions, wherein the program instructions, when executed, control the device where the storage medium is located to perform the above-described method.

[0068] In summary, this invention constructs a three-dimensional model of the target area by combining historical and time-series images from the LT-1 satellite, and extracts information on crack areas and their changes, thus achieving high-precision monitoring of deformation on ordinary highway slopes. This method provides more accurate deformation data, improving the accuracy and reliability of monitoring compared to traditional methods. Utilizing time-series images from the LT-1 satellite, this invention can update displacement and morphological changes in the target area in real time, obtaining the latest deformation time-series information. This real-time monitoring capability enables management departments to respond promptly to slope deformation, improving the timeliness of geological disaster response. It automatically extracts crack information through image recognition models, ensuring the accuracy of these models even without the need for precise annotation of sample images. This reduces labor intensity and human error, improving the efficiency and accuracy of crack analysis. By combining first and second deformation time-series information, and through the synergy of the above technical solutions, it can not only predict the future deformation trend and type of slopes, providing a scientific basis for preventing and mitigating geological disasters, but also reduce reliance on ground sensors and manual inspections, lowering monitoring costs while improving monitoring efficiency. Especially in remote or inaccessible areas, this invention not only focuses on crack deformation but also incorporates displacement and morphological change information. Through data fusion technology, it provides a more comprehensive deformation analysis. The method of this invention is applicable to slope deformation monitoring under various environmental conditions, unrestricted by weather, terrain, or other natural conditions, and has strong environmental adaptability.

[0069] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0070] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0071] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for time-series monitoring of deformation of ordinary highway slopes based on LT-1 satellite, characterized in that, The method is as follows: Historical images of the target area are acquired, the historical images are preprocessed, a fixed object outside the target area is selected as a reference object, and the reference object is used as the origin of the coordinate system. A three-dimensional model of the target area is constructed based on the historical images and the origin of the coordinate system. Each satellite image in the time-series images acquired within a preset time period is input into the image recognition model to obtain the recognition image corresponding to the satellite image. Based on each recognition image, the crack region in the target region is extracted, and the crack information corresponding to the crack region in each satellite image is obtained. The crack change is obtained through all the crack information corresponding to the time-series images, and the first deformation time-series information is obtained based on the crack change. Each satellite image in the time series is updated to the three-dimensional model in chronological order, and the position and shape information of the slope in the target area in each satellite image are obtained. Based on the position and shape information of the slope corresponding to each satellite image, the displacement and shape change information of the target area within the preset time period are obtained. Based on the displacement and shape change information, the second deformation time series information is obtained. The slope deformation prediction results are obtained based on the first deformation time series information, the second deformation time series information, and the deformation prediction model. The image recognition model is a model trained with sample images and corresponding first label information, wherein the first label information consists of slope pixel label information corresponding to some slope areas and crack pixel label information corresponding to some crack areas in the sample image. The deformation prediction model is a model trained on historical deformation data, which includes displacement deformation, slope morphology changes, crack deformation and corresponding deformation types, as well as displacement deformation, slope morphology changes and crack deformation during the entire deformation process of the deformation type.

2. The method for time-series monitoring of ordinary highway slope deformation based on LT-1 satellite according to claim 1, characterized in that, The acquisition of the crack information includes: Each satellite image in the time-series images acquired within the preset time period is preprocessed and input into the image recognition model to obtain the recognition image corresponding to each satellite image; wherein, the recognition image is a binary image, and is an image that distinguishes the crack area and the slope area in the corresponding target area by pixel value, and the satellite image is a three-dimensional image; Crack regions are extracted based on the crack probability values ​​corresponding to cracks in the identified image. Each crack region is divided into multiple sub-segments according to a set vertical distance. The first and second tangents on both sides of each sub-segment are obtained. A first and second angle are calculated between each adjacent sub-segment and the first and second tangents on the same side of the sub-segment. The first and second angles are angles within the crack region. If at least one of the first and second angles between at least one adjacent sub-segment and the sub-segment is less than a set angle, the sub-segment is considered a crack segment. When both the first included angle and the second included angle are greater than or equal to a set angle, at least one of the adjacent sub-segments is taken as the target sub-segment, and the target sub-segment and the sub-segment are merged into a new sub-segment. The two sides of the first included angle are connected to the two intersection points of the corresponding sub-segment and used as the third tangent of the new sub-segment. Similarly, the fourth tangent on the other side of the new sub-segment is obtained, and this step is repeated to obtain multiple crack segments of the crack region. The crack included angle between adjacent crack segments is greater than or equal to the set angle. The crack region in the recognition image, the crack segment corresponding to each crack region, and the size and position information of each crack segment are used as the crack information.

3. The method for time-series monitoring of ordinary highway slope deformation based on LT-1 satellite according to claim 1, characterized in that, The acquisition of the first deformation timing information of the target region includes: Crack change information is obtained by comparing the crack segment in the crack information corresponding to each satellite image in the time-series image with the crack information of the crack segment at the same position in the crack information corresponding to the adjacent satellite images in the time sequence, wherein the acquisition time of the adjacent satellite images is later than the acquisition time of the satellite images. Based on the crack change information corresponding to each satellite image and the adjacent satellite images, the first deformation time-series information of the target area in the time-series image within the preset time period is obtained, wherein the crack change information includes size change and depth change.

4. The method for time-series monitoring of ordinary highway slope deformation based on LT-1 satellite according to claim 1, characterized in that, The training of the image recognition model includes: The image recognition model is created by the model creation unit, and a sample image is read from the storage unit. The sample image includes first label information, which is a pre-labeled label. The sample image and the first label information are used as first training data to train the image recognition model. During the training iteration, the average moving weight of the prediction result corresponding to each pixel in the sample image is periodically calculated, and the average moving weight is used to replace the prediction result obtained in the current iteration. Second label information is set for second pixels whose prediction result is greater than a set threshold, wherein the second pixel is a pixel that is not labeled with the pre-labeled label, and the pixel that includes the pre-labeled label is used as the first pixel. At least two second pixels adjacent to the first pixels are obtained, and at least two first pixels are used as reference pixels. The first label information corresponding to the reference pixels is the same and located on both sides of the second pixel. The second label information corresponding to the second pixel is compared with the first label information corresponding to the reference pixel. When the second label information is the same as the first label information corresponding to the reference pixel, the second pixel and the corresponding second label information are added as first learning sample pixels to the second training data. Otherwise, the difference features between the second pixel and the reference pixels are extracted again and enhanced. The image recognition model is then retrained using the second training data.

5. The method for time-series monitoring of ordinary highway slope deformation based on LT-1 satellite according to claim 1, characterized in that, The acquisition of the second deformation timing information includes: Each satellite image obtained within the preset time period is preprocessed and then sequentially updated to the three-dimensional model. The position and shape information of the slope in the temporally adjacent satellite images are compared to obtain displacement change information and slope shape change information. The displacement change information and slope shape change information are used as the second deformation time series information.

6. The method for time-series monitoring of ordinary highway slope deformation based on LT-1 satellite according to claim 1, characterized in that, Obtaining the slope deformation prediction results includes: The first deformation time series information and the second deformation time series information are fused according to time series to obtain slope deformation time series information, and the slope deformation time series information is input into the deformation prediction model to obtain the slope deformation prediction result. The slope deformation prediction result includes the future deformation development trend and slope deformation type of the target area.

7. A method for time-series monitoring of ordinary highway slope deformation based on LT-1 satellite according to claim 6, characterized in that, The deformation development trend of the target area includes the displacement change trend, morphological change trend, and crack change trend of the slope over time.

8. A time-series monitoring system for ordinary highway slope deformation based on TL-1 satellite, the system being used to implement the time-series monitoring method for ordinary highway slope deformation based on LT-1 satellite as described in any one of claims 1-7, characterized in that, The system includes: The model creation unit is used to acquire historical images of the target area, preprocess the historical images, select a fixed object outside the target area as a reference object, and use the reference object as the origin of the coordinate system to construct a three-dimensional model of the target area based on the historical images and the origin of the coordinate system. The computing unit is configured to: input each satellite image in the time-series images acquired within the preset time period into an image recognition model to obtain a recognition image corresponding to the satellite image; extract the crack region in the target region based on each recognition image and obtain crack information corresponding to the crack region in each satellite image; obtain crack changes through all the crack information corresponding to the time-series images and obtain first deformation time-series information based on the crack changes; update each satellite image in the time-series images to the three-dimensional model in chronological order and obtain the position and shape information of the slope in the target region in each satellite image; obtain the displacement and shape change information of the target region within the preset time period based on the position and shape information of the slope corresponding to each satellite image; and obtain second deformation time-series information based on the displacement and shape change information. The prediction unit is used to obtain the slope deformation prediction result based on the first deformation time series information, the second deformation time series information, and the deformation prediction model.

9. A computer storage medium, characterized in that, The storage medium stores program instructions, wherein when the program instructions are executed, the device where the storage medium is located executes the time-series monitoring method for ordinary highway slope deformation based on LT-1 satellite as described in any one of claims 1-7.

Images