Sand dune movement speed change remote sensing detection method and system, electronic equipment and storage medium

Abstract

The invention provides a sand dune movement speed change remote sensing detection method which can be applied to the technical field of remote sensing image processing. The method comprises the following steps: extracting a sand dune edge from a sand dune remote sensing observation time sequence image to obtain a sand dune edge image sequence, and preprocessing the sand dune edge image sequence to obtain a sand ridge line image sequence; constructing a velocity vector field for two adjacent frames of sand ridge line image data in the sand ridge line image sequence by using a dense optical flow method to obtain a sand ridge line moving velocity field space-time cube array; and performing pixel-by-pixel time sequence analysis on the sand ridge line moving speed field time-space cube array by using the exponential weighted moving average control chart to obtain a time sequence rate abnormal point judgment result, and overlapping the time sequence rate abnormal point judgment result on a sand dune remote sensing observation time sequence image for identification to obtain a sand dune moving speed change detection result. The invention further provides a sand dune movement speed change remote sensing detection system, electronic equipment and a storage medium.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensing image processing, and specifically relates to a method and system for remotely detecting the rapid change of dune movement, an electronic device, and a storage medium. Background Art

[0002] The rapid change of dune movement means that the speed of the dune changes violently during the movement process. At present, the monitoring of dune movement mainly relies on ground measurement techniques and remote sensing image techniques. Among them, ground measurement techniques mainly use methods such as the stake insertion method, total station, theodolite, and real-time kinematic differential GPS. Its advantage lies in high precision, but due to the harsh field environment and work efficiency limitations, there are problems such as large workload, high cost, and small scope; while remote sensing monitoring mainly uses satellite remote sensing images and unmanned aerial vehicle aerial images, which have the advantages of quickly obtaining dune movement parameters, large scope, and a large number of monitored dunes. However, the measurement of dune movement mainly still uses the method of visual interpretation, that is, using geographic information processing software to manually outline the dune contours in the image and select moving feature points (for example, the five-point measurement method, centroid measurement method, etc.) for measurement to show the macroscopic movement of the dunes.

[0003] In the existing technical solutions, the extraction of dune movement parameters from remote sensing images mainly still uses visual interpretation to manually calibrate feature points using geographic information processing software for measurement, and most studies only select some representative feature points to describe the macroscopic movement of the dunes, making it difficult to characterize the detailed changes in the movement of the entire mobile dune. If more detailed rapid change detection of dune movement and morphological changes is required, then more time-series satellite images with shorter time intervals are needed; and the method of manually annotating and measuring dune movement will undoubtedly consume a lot of time; in addition, there is a lack of technical solutions for the rapid change detection of dune movement in the existing technical solutions. Summary of the Invention

[0004] In view of the above problems, the present invention provides a method and system for remotely detecting the rapid change of dune movement, an electronic device, and a storage medium, which are used to solve at least one of the existing technical problems.

[0005] According to a first aspect of the present invention, a method for remotely detecting the rapid change of dune movement is provided, including:

[0006] Using a deep learning method to extract the dune edge from the time-series images of dune remote sensing observations, obtaining a sequence of dune edge images, and preprocessing the sequence of dune edge images to obtain a sequence of sand ridge line images;

[0007] Using the dense optical flow method to construct a velocity vector field for the sand ridge line image data of two adjacent frames in the sequence of sand ridge line images, obtaining a spatio-temporal cube array of the sand ridge line movement speed field;

[0008] Perform pixel-by-pixel time series analysis on the spatio-temporal cube array of the sand ridge line movement speed field using an exponentially weighted moving average control chart to obtain the judgment results of time series rate anomaly points, and identify the position-time by superimposing the judgment results of the time series rate anomaly points on the time series image of the sand dune remote sensing observation, so as to obtain the sand dune movement speed change detection results.

[0009] According to an embodiment of the present invention, the preprocessing of the sand dune edge image sequence to obtain the sand ridge line image sequence includes:

[0010] Use the erosion filtering method to filter the noise of the sand dune edge image sequence to obtain the denoised sand dune edge image sequence;

[0011] Based on the sand dune edge image sequence, use the bilinear interpolation method to resample the denoised sand dune edge image sequence to obtain the sand ridge line image sequence.

[0012] According to an embodiment of the present invention, the construction of the velocity vector field for the sand ridge line image data of two adjacent frames in the sand ridge line image sequence using the dense optical flow method to obtain the spatio-temporal cube array of the sand ridge line movement speed field includes:

[0013] Use the Farneback dense optical flow model to calculate the horizontal displacement and vertical displacement of each pixel in the sand ridge line image data of two adjacent frames to obtain the time series of the displacement vector field;

[0014] Stitch the time series of the displacement vector fields of all adjacent two-frame sand ridge line image data to obtain the stitching result;

[0015] Based on the stitching result, calculate the displacement vector of each adjacent two-frame sand ridge line image data and the time interval of each adjacent two-frame sand ridge line image data to obtain the spatio-temporal cube array of the sand ridge line movement speed field.

[0016] According to an embodiment of the present invention, the spatio-temporal cube array of the sand ridge line movement speed field includes time axis data, image horizontal axis data, image vertical axis data, pixel horizontal displacement data, pixel vertical displacement data, pixel horizontal speed data, and pixel vertical speed data, and the spatio-temporal cube array of the sand ridge line movement speed field is used to characterize the speed information of the pixels in the sand ridge line image sequence.

[0017] According to an embodiment of the present invention, the pixel-by-pixel time series analysis of the spatio-temporal cube array of the sand ridge line movement speed field using an exponentially weighted moving average control chart to obtain the judgment results of time series rate anomaly points includes:

[0018] Extract the displacement sequence and movement rate of each image coordinate on the sand ridge line from the spatio-temporal cube array of the sand ridge line movement speed field to obtain the spatio-temporal sequence of the sand ridge line movement rate;

[0019] Use an exponentially weighted moving average control chart to discriminate time - series rate outliers from the spatio - temporal sequence of the sand ridge line movement rate, and obtain the judgment result of time - series rate outliers.

[0020] According to an embodiment of the present invention, the above - mentioned extraction of the displacement sequence and movement rate of each image coordinate on the sand ridge line from the spatio - temporal cube array of the sand ridge line movement speed field, and obtaining the spatio - temporal sequence of the sand ridge line movement rate includes:

[0021] Perform pixel - by - pixel labeling on the spatio - temporal cube array of the sand ridge line movement speed field at each moment to obtain the pixel - labeled points of the sand ridge line at each moment;

[0022] Add the horizontal axis velocity vector and the vertical axis velocity vector corresponding to the pixel - labeled points of the sand ridge line at each moment, and then perform a modulus operation to obtain the spatio - temporal sequence of the sand ridge line movement rate.

[0023] According to an embodiment of the present invention, the above - mentioned use of an exponentially weighted moving average control chart to discriminate time - series rate outliers from the spatio - temporal sequence of the sand ridge line movement rate, and obtaining the judgment result of time - series rate outliers includes:

[0024] Use the spatio - temporal sequence of the sand ridge line movement rate to construct a recurrence formula for the statistic of the exponentially weighted moving average control chart, and expand the recurrence formula of the statistic into the form of an exponentially weighted average of all time - rate sequences to obtain the expansion of the statistic;

[0025] Based on the maximum likelihood estimation method, construct an estimate of the standard deviation of the statistic, and use the estimate of the standard deviation of the statistic to operate on the expansion of the statistic to obtain the upper and lower control limits and the center line of the exponentially weighted moving average control chart;

[0026] Use the upper and lower control limits and the center line of the exponentially weighted moving average control chart to perform pixel - by - pixel discrimination of time - series rate outliers on the spatio - temporal sequence of the sand ridge line movement rate, and determine the pixels whose time - series rate exceeds the upper and lower control limits of the exponentially weighted moving average control chart as time - series rate outliers.

[0027] The second aspect of the present invention provides a remote - sensing detection system for the changing speed of sand dune movement, including:

[0028] A remote - sensing image pre - processing module, which is used to extract the sand dune edge from the time - series images of sand dune remote sensing observations by using deep - learning methods to obtain a sequence of sand dune edge images, and perform pre - processing on the sequence of sand dune edge images to obtain a sequence of sand ridge line images;

[0029] A velocity vector field construction module, which is used to construct a velocity vector field for the sand ridge line image data of two adjacent frames in the sequence of sand ridge line images by using the dense optical flow method to obtain a spatio - temporal cube array of the sand ridge line movement speed field;

[0030] The dune rapid change detection module is used to perform pixel-by-pixel time series analysis on the spatio-temporal cube array of the sand ridge line movement speed field by using an exponentially weighted moving average control chart, obtain the judgment result of time series rate abnormal points, and identify the position-time by superimposing the judgment result of time series rate abnormal points on the time series image of dune remote sensing observation, so as to obtain the dune movement rapid change detection result.

[0031] The third aspect of the present invention provides an electronic device, including: one or more processors; a memory for storing one or more computer programs, wherein the above one or more processors execute the above one or more computer programs to implement the steps of the above method.

[0032] The fourth aspect of the present invention further provides a computer-readable storage medium, on which a computer program or instruction is stored, and when the computer program or instruction is executed by a processor, the steps of the above method are implemented.

[0033] The dune movement rapid change remote sensing detection method provided by the present invention can realize the automatic estimation of dune movement in remote sensing images and can estimate dune movement information from a microscopic angle by constructing the dune movement state through the dense optical flow method; in addition, the dune movement rapid change remote sensing detection method provided by the present invention can effectively detect the rapid changes of dune movement in a remote sensing image sequence by using a simple and efficient detection method based on the optical flow method and the exponentially weighted moving average control chart. Description of the Drawings

[0034] Through the following description of the embodiments of the present invention with reference to the drawings, the above content and other objects, features and advantages of the present invention will become clearer. In the drawings:

[0035] Figure 1 is an application scenario diagram of the dune movement rapid change remote sensing detection method according to an embodiment of the present invention;

[0036] Figure 2 is a flowchart of the dune movement rapid change remote sensing detection method according to an embodiment of the present invention;

[0037] Figure 3 is a data processing framework diagram of the dune movement rapid change detection method according to an embodiment of the present invention;

[0038] Figure 4 is a schematic diagram of the optical flow field established between two adjacent frames according to an embodiment of the present invention;

[0039] Figure 5 is a schematic diagram of the dune movement rapid change detection result according to an embodiment of the present invention;

[0040] Figure 6 is a schematic diagram of the basic form of the quality control chart according to an embodiment of the present invention;

[0041] Figure 7 It is a schematic structural diagram of a remote sensing detection system for rapid change of sand dune movement according to an embodiment of the present invention;

[0042] Figure 8 It is a schematic diagram of the actual scenes of two deserts, Sand1 and Sand2, according to an embodiment of the present invention and the corresponding schematic diagram of sand ridge line extraction;

[0043] Figure 9 It is an input image and the corresponding output image of rapid change detection of Sand1 according to an embodiment of the present invention;

[0044] Figure 10 It is a schematic diagram of the rapid change detection results of Sand1 during the years 2020 - 2021 according to an embodiment of the present invention;

[0045] Figure 11 It is a schematic diagram of the rapid change detection results of Sand1 during the years 2010 - 2014 according to an embodiment of the present invention;

[0046] Figure 12 It is an input image and the corresponding output image of rapid change detection of Sand2 according to an embodiment of the present invention;

[0047] Figure 13 It is a schematic diagram of the rapid change detection results of Sand2 during the years 2010 - 2014 according to an embodiment of the present invention;

[0048] Figure 14 It is a block diagram of an electronic device suitable for implementing the method for remote sensing detection of rapid change of sand dune movement according to an embodiment of the present invention. Detailed implementation manners

[0049] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. In the following detailed description, for the sake of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present invention. However, obviously, one or more embodiments can also be implemented without these specific details. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessarily obscuring the concepts of the present invention.

[0050] The terms used herein are merely for describing specific embodiments and are not intended to limit the present invention. The terms "including", "comprising", etc. used herein indicate the presence of the described features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0051] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those of ordinary skill in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0052] In cases where expressions similar to "at least one of A, B, and C, etc." are used, generally, it should be interpreted according to the meaning commonly understood by those of ordinary skill in the art (for example, "a system having at least one of A, B, and C" should include, but not be limited to, a system having only A, only B, only C, having A and B, having A and C, having B and C, and / or having A, B, and C, etc.).

[0053] Figure 1 It is an application scenario diagram of the remote sensing detection method for the variable speed of sand dune movement according to an embodiment of the present invention.

[0054] As Figure 1 shown, the application scenario 100 according to this embodiment may include the field of remote sensing image processing technology. The network 104 is used to provide a medium for communication links between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc.

[0055] Users can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications may be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only for example).

[0056] The first terminal device 101, the second terminal device 102, and the third terminal device 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop portable computers, and desktop computers, etc.

[0057] The server 105 may be a server providing various services, such as a background management server that supports the websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (only for example). The background management server may analyze and process data such as user requests received, and feedback the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0058] It should be noted that the method for remotely sensing and detecting the variable speed of sand dune movement provided by the embodiments of the present invention can generally be executed by the server 105. Correspondingly, the system for remotely sensing and detecting the variable speed of sand dune movement provided by the embodiments of the present invention can generally be set in the server 105. The method for remotely sensing and detecting the variable speed of sand dune movement provided by the embodiments of the present invention can also be executed by a server or a server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or the server 105. Correspondingly, the system for remotely sensing and detecting the variable speed of sand dune movement provided by the embodiments of the present invention can also be set in a server or a server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or the server 105.

[0059] It should be understood that Figure 1 the numbers of the terminal devices, networks, and servers in

[0060] are merely illustrative. According to the implementation requirements, there can be any number of terminal devices, networks, and servers. Figure 1 The following will be based on Figures 2 - 6 the described scenario, and will describe in detail the method for remotely sensing and detecting the variable speed of sand dune movement of the disclosed embodiments through

[0061] Figure 2 is a flowchart of the method for remotely sensing and detecting the variable speed of sand dune movement according to the embodiments of the present invention.

[0062] As Figure 2 shown, the above method for remotely sensing and detecting the variable speed of sand dune movement includes operations S210 to S230.

[0063] In operation S210, the sand dune edge is extracted from the sequential images of remote sensing observations of the sand dune by using a deep learning method to obtain a sequence of sand dune edge images, and the sequence of sand dune edge images is preprocessed to obtain a sequence of sand ridge line images.

[0064] Since the size of the sequential images of remote sensing observations of the sand dune is large, before detecting the variable speed of the sand dune movement, it is necessary to process the sequential images of remote sensing observations of the sand dune: first, the sand dune edge is extracted by using a deep learning method; then, the extracted sequence of sand dune edge images is preprocessed, such as operations such as filtering and downsampling.

[0065] In operation S220, a velocity vector field is constructed for the sand ridge line image data of two adjacent frames in the sequence of sand ridge line images by using the dense optical flow method to obtain a spatio-temporal cube array of the sand ridge line movement speed field.

[0066] For the above dense optical flow method, optionally, the Farneback dense optical flow method can be used.

[0067] In operation S230, a pixel-by-pixel time series analysis is performed on the spatio-temporal cube array of the sand ridge line movement speed field using an exponentially weighted moving average control chart to obtain the judgment result of time series rate anomaly points, and the judgment result of time series rate anomaly points is superimposed on the time series image of dune remote sensing observation to identify the position-time, thereby obtaining the dune movement speed change detection result.

[0068] The dune movement speed change remote sensing detection method provided by the present invention can realize the automatic estimation of dune movement in remote sensing images and can estimate dune movement information from a microscopic perspective by constructing the dune movement state through the dense optical flow method; in addition, the dune movement speed change remote sensing detection method provided by the present invention can effectively detect the rapid changes in dune movement in a remote sensing image sequence by using a simple and efficient detection method based on the optical flow method and the exponentially weighted moving average control chart.

[0069] The following further elaborates in detail on the above-mentioned dune movement speed change detection method provided by the present invention through specific embodiments in combination with the attached Figures 3 - 5 drawings.

[0070] Figure 3 FIG. is a data processing framework diagram of the dune movement speed change detection method according to an embodiment of the present invention.

[0071] Figure 4 FIG. is a schematic diagram of the optical flow field established between two adjacent frames according to an embodiment of the present invention.

[0072] Figure 5 FIG. is a schematic diagram of the dune movement speed change detection result according to an embodiment of the present invention.

[0073] Taking a certain desert area as the research object, the optical flow method is used to construct a time series of the speed field for the sand ridge line data of the extracted dunes, and the speed anomaly points are detected through a time series anomaly detection method, thereby realizing the automatic detection and analysis of dune movement speed changes. First, the dune images and their sand ridge line extraction images in a specific time period, such as from 2010 to 2022, are obtained. Subsequently, the dense optical flow method is used to model the optical flow changes between the frames of the sand ridge line images to describe the movement state of the dunes. Finally, the rate anomaly points are detected through the EWMA (exponentially weighted moving average) control chart to complete the monitoring of rate changes. The innovation of the method provided by the present invention lies in the combination of the dense optical flow method and the EWMA control chart, which realizes the automatic detection of dune movement speed changes from a microscopic perspective, can quickly and accurately identify the sudden acceleration and deceleration of dune movement speed and the appearance and disappearance of the sand ridge line, and provides objective and efficient repeatable analysis results.

[0074] As Figure 3 shown, Figure 3On the left is the process of establishing the optical flow field, and on the right is the process of detecting velocity field anomalies. First, the sequence of dune edge images extracted from the desert remote sensing observation image sequence using deep learning is used as the input. Currently, the algorithms for directly extracting the sand ridge line are not yet perfect, mainly targeting the extraction of dune edges. Since the difference between the dunes and the desert background is small, there are many noise points and discontinuity points during the extraction process. Therefore, the erosion filtering method is used to remove some noise points, making the extracted sand ridge line smoother and having less background noise. In addition, since remote sensing images often have too large pixel widths, such as Figure 3 the input image in it has a pixel width of 2357 2268, while the commonly used optical flow method is often developed based on computer vision images, and the applicable image sizes are often small. After experiments, the image is resampled to 512 512. The discontinuity points in the dune edge extraction are reduced, the sand ridge line is clearer and smoother, and the effect of constructing the velocity field is optimal. At the same time, the resampling operation can reduce the computational amount of subsequent time series analysis and greatly improve the speed change detection efficiency. Finally, the sequence of sand ridge line images obtained after preprocessing is input into the Farneback dense optical flow method to establish the optical flow field. The Farneback dense optical flow method is based on the optical flow field modeling between two frames, so the input is a pair of sand ridge line images, such as Figure 4 the optical flow field established by adjacent double frames shown, where Figure 4 the bottom map of is the sand ridge line image of the latter frame.

[0075] During the process of establishing the optical flow field, the present invention mainly analyzes from the spatial domain and studies the spatial change of the optical flow between two-frame images. Specifically, the optical flow field is repeatedly established for every two frames of images in chronological order and they are superimposed together to form a spatio-temporal cube array of the sand ridge line moving velocity field. Subsequently, the spatio-temporal cube array of the sand ridge line moving velocity field is transferred to the time domain, and each pixel on the sand ridge line is analyzed step by step for each time.

[0076] The present invention provides a more robust EWMA control chart time series anomaly detection method, which can effectively detect the velocity change points of dune movement, such as Figure 5 shown, and the detection results are displayed in the bottom map of the original image, corresponding to Figure 4 the optical flow field and the sand ridge line extraction map in. The method provided by the present invention has the advantages of being simple and efficient, having high detection accuracy, not requiring a large amount of training data, and being robust. It can quickly process large-scale time series data and shows excellent detection ability when capturing the velocity change points of dune movement. Especially in the case of small changes, it can still effectively identify anomalies. Compared with some complex machine learning models, the method provided by the present invention does not rely on large-scale labeled data, reducing the data preparation and training costs. At the same time, the method provided by the present invention can still stably output accurate results in the presence of noise or imperfect data.

[0077] According to an embodiment of the present invention, the above preprocessing of the dune edge image sequence to obtain the sand ridge line image sequence includes: acquiring the time-series images of dune remote sensing observations, extracting the dune edge from the time-series images of dune remote sensing observations to obtain the dune edge image sequence; using the erosion filtering method to filter the noise of the dune edge image sequence to obtain the denoised dune edge image sequence; based on the dune edge image sequence, using the bilinear interpolation method to resample the denoised dune edge image sequence to obtain the sand ridge line image sequence.

[0078] In an example of the present invention, a time-series image sequence of dune images in the same area is obtained from an open-source database (such as Google Earth), and the dune edge image sequence is extracted by using deep learning methods. Since there are many noise points and discontinuity points in the extracted dune edge image sequence, which is not conducive to the subsequent estimation of the motion field, the erosion filtering method is used to filter out the noise points. At the same time, to solve the problems of the extraction effect and efficiency of the optical flow field caused by the discontinuity points of the dune edge and the excessive pixel width of the image, the bilinear interpolation method is used to resample the denoised dune edge image sequence. After multiple experiments, the image is resampled to 512 A size of 512×512 is more conducive to the extraction of the sand ridge line and the subsequent optical flow field processing, and finally the sand ridge line image sequence is obtained.

[0079] According to an embodiment of the present invention, the above construction of the velocity vector field for the sand ridge line image data of two adjacent frames in the sand ridge line image sequence by using the dense optical flow method to obtain the spatio-temporal cube array of the sand ridge line moving speed field includes: using the Farneback dense optical flow model to calculate the horizontal displacement and vertical displacement of each pixel in the sand ridge line image data of two adjacent frames to obtain the time series of the displacement vector field; splicing the time series of the displacement vector fields of all the sand ridge line image data of two adjacent frames to obtain the splicing result; based on the splicing result, performing an operation on the displacement vector of each adjacent two-frame sand ridge line image data and the time interval between each adjacent two-frame sand ridge line image data to obtain the spatio-temporal cube array of the sand ridge line moving speed field.

[0080] According to an embodiment of the present invention, the above spatio-temporal cube array of the sand ridge line moving speed field includes time-axis data, image horizontal-axis data, image vertical-axis data, pixel horizontal displacement data, pixel vertical displacement data, pixel horizontal speed data, and pixel vertical speed data, and the spatio-temporal cube array of the sand ridge line moving speed field is used to characterize the speed information of the pixels in the sand ridge line image sequence.

[0081] Optical flow refers to the instantaneous velocity of the pixel motion of a moving object in space on the observation imaging plane. The optical flow method is a way to calculate the motion information of an object between adjacent frames by using the change of pixels in the time domain of an image sequence and the correlation between adjacent frames to find the corresponding relationship between the previous frame and the current frame.

[0082] The implementation of the optical flow algorithm needs to meet two basic assumptions: the brightness of the same object remains unchanged when it moves in two adjacent frames of images; the movement of the target object is relatively small, that is, the object undergoes small displacement movement. The basic constraint equation of optical flow is proposed under the basic assumption of grayscale conservation, which can be understood as that the pixel grayscale remains unchanged after the target object moves.

[0083] The dense optical flow method is an image registration method that performs point-by-point matching for an image or a specified area. Different from the sparse optical flow method (such as the LK algorithm) which usually needs to specify a set of points for tracking, it needs to calculate the offset of all points on the image to form a dense optical flow field, and then perform pixel-level image registration through this dense optical flow field. The HS algorithm and most optical flow methods based on region matching belong to the category of dense optical flow. The Farneback algorithm is a dense optical flow method based on polynomial expansion. The accuracy of the Farneback optical flow method is higher than that of the sparse optical flow method LK algorithm, and it has a faster calculation speed compared with the similar dense optical flow method HS algorithm. Therefore, the present invention uses the Farneback dense optical flow method to process the involved image sequence.

[0084] In the specific implementation manner of the present invention, the optical flow estimation is performed on two adjacent frames of sand ridge line images in the sand ridge line image sequence in the order of time using the Farneback optical flow method. The parameter design of the Farneback optical flow model is as follows: construct a pyramid layer with a scaling ratio of 0.5 for 3 layers, use a mean window with a size of 15 for matching, the number of executions of the iterative algorithm on each level of the pyramid is 5, the neighborhood size of the pixel range searched is 1.2 and its standard deviation is 0. The model output is the horizontal and vertical displacement of each pixel in two adjacent frames of images. Input all the image sequences into this model in the way of every two frames in turn, and a time series of displacement vector fields can be obtained. Stitch the obtained output results, and divide the displacement vector by the time interval between two frames to obtain the speed. Finally, a spatio-temporal cube array of the sand ridge line movement speed field is obtained, which can represent the speed of each pixel in the sand ridge line image sequence. The dimensions of the spatio-temporal cube array include: time axis t, horizontal axis x of the image, vertical axis y of the image, horizontal displacement sx of the pixel, vertical displacement sy of the pixel, horizontal speed vx of the pixel, and vertical speed vy of the pixel.

[0085] According to an embodiment of the present invention, the above-mentioned time-series analysis of each pixel of the spatio-temporal cube array of the sand ridge line movement speed field using an exponentially weighted moving average control chart to obtain the judgment result of time-series rate anomaly points includes: extracting the displacement sequence and movement rate of each image coordinate on the sand ridge line from the spatio-temporal cube array of the sand ridge line movement speed field to obtain the spatio-temporal sequence of the sand ridge line movement rate; using the exponentially weighted moving average control chart to discriminate the time-series rate anomaly points of the spatio-temporal sequence of the sand ridge line movement rate to obtain the judgment result of the time-series rate anomaly points.

[0086] According to an embodiment of the present invention, the above-mentioned extraction of the displacement sequence and movement rate of each image coordinate on the sand ridge line from the spatio-temporal cube array of the sand ridge line movement speed field to obtain the spatio-temporal sequence of the sand ridge line movement rate includes: performing pixel-by-pixel marking on the spatio-temporal cube array of the sand ridge line movement speed field at each moment to obtain the sand ridge line pixel marking points at each moment; adding the horizontal axis velocity vector and the vertical axis velocity vector corresponding to the sand ridge line pixel marking points at each moment and then performing a modulus operation to obtain the spatio-temporal sequence of the sand ridge line movement rate.

[0087] Extract the displacement sequence of each image coordinate on the sand ridge line from the spatio-temporal cube array of the sand ridge line movement speed field, and at the same time extract the movement rate of each image coordinate of the sand ridge line to obtain the time-series data of the sand ridge line movement rate , use the EWMA control chart to discriminate the rate anomaly points of the time-series data of the sand ridge line movement rate, and superimpose the rate anomaly point discrimination result on the time-series image of the sand dune remote sensing observation for identification to complete the detection of the sand dune movement speed change.

[0088] In an embodiment of the present invention, the displacement sequence of each image coordinate on the sand ridge line is extracted, and the displacement sequence represents the movement of each image coordinate on the sand ridge line. The sand ridge line in the sand ridge line image at the first moment is marked pixel by pixel to obtain the sand ridge line pixel marking points (a total of pixel marking points, marked as . For the l-th sand ridge line pixel marking point , the displacement sequence is expressed as , and the relationship between two adjacent moments in the displacement sequence is and relationship); calculate the rate of the sand ridge line pixel marking point at the pixel position at the moment, where the calculation method of the value is to convert the velocity vector into a rate scalar , that is, add the horizontal and vertical axis velocity vectors of the pixels perpendicular to each other and then extract the modulus value. If the value at the pixel position is 0, then at the Search within the 3×3 window of pixel coordinates Replace with the pixel coordinates having the maximum value to obtain the time series data of the moving rate of the sand ridge line Use the EWMA (Exponentially Weighted Moving Average) control chart to perform outlier detection on the time series data of the moving rate of the sand ridge line to obtain the time series data of the rapid change points of the moving rate of the sand ridge line including the pixel marker points of the sand ridge line with rapid change points of the rate pixel coordinates and time .

[0089] According to an embodiment of the present invention, the above-mentioned use of the exponentially weighted moving average control chart to perform time series rate outlier discrimination on the spatio-temporal series of the moving rate of the sand ridge line, and the obtained time series rate outlier judgment results include: constructing a recurrence formula for the statistic of the exponentially weighted moving average control chart using the spatio-temporal series of the moving rate of the sand ridge line, and expanding the recurrence formula for the statistic into the form of the exponentially weighted average of all time rate series to obtain the expansion formula of the statistic; based on the maximum likelihood estimation method, constructing an estimation of the standard deviation of the statistic, and using the estimation of the standard deviation of the statistic to perform operations on the expansion formula of the statistic to obtain the upper and lower control limits and the center line of the exponentially weighted moving average control chart; using the upper and lower control limits and the center line of the exponentially weighted moving average control chart to perform pixel-by-pixel time series rate outlier discrimination on the spatio-temporal series of the moving rate of the sand ridge line, and determining the pixels whose time series rate exceeds the upper and lower control limits of the exponentially weighted moving average control chart as time series rate outliers.

[0090] To facilitate the exponentially weighted moving average control chart used in the present invention, the exponentially weighted moving average control chart is described below.

[0091] The quality control chart shows that there are variations in each data analysis and monitoring method, and they are affected by time and space. Even a set of analysis results obtained under ideal conditions will have certain random errors. When a certain result exceeds the allowable range of random errors, using the method of mathematical statistics, it can be judged that this result is abnormal and untrustworthy.

[0092] Figure 6 is a schematic diagram of the basic form of the quality control chart according to an embodiment of the present invention.

[0093] The quality control chart is a chart used to analyze the change of a process over time, and is a chart with control limits used to analyze and judge whether a sequence is in a stable state. The basic form of the quality control chart is as Figure 6As shown in the figure. The abscissa in the figure is the time point (Time Points), and the ordinate is the value of a certain statistic of the process (Statistics). There are three lines on the quality control chart: the upper control limit (Upper Control Limit, UCL), the central line (Central Line, CL), and the lower control limit (Lower Control Limit, LCL). The central line is the average level of the process statistic value. The area [LCL, UCL] between the upper / lower control limits is the acceptance region, and the area outside this interval is called the rejection region. By comparing the statistic value of the process with the upper and lower control limits, if the statistic value enters the rejection region, it is judged that the process is out of control, that is, there is an abnormality.

[0094] The quality control chart is the application of significance testing in statistics to process stability control, and its principle is statistical hypothesis testing. Each quality control chart is a statistical hypothesis test. Under normal conditions, the process is relatively stable. When only random factors act, according to the central limit theorem, the overall distribution of these random errors is a normal distribution or an approximately normal distribution. Specifically, the output of a process has its statistical characteristics, and the stability of the process can be monitored by judging whether these statistical characteristics are stable. For the output of the process select a statistic (Statistics) , and at the same time select a significance level (such as 0.001), then there is a corresponding rejection region. Under normal conditions, that is, when the process is stable, the probability of falling within the rejection region is a small probability . Since it is generally considered that small probability events will not occur, so if falls within the rejection region, it is considered that the statistic value of the process is abnormal, and it is determined that there is an abnormality in the process at the corresponding moment.

[0095] The upper / lower control limits (UCL / LCL) and the central line (CL) of the quality control chart are determined by formula (1) as follows:

[0096] (1),

[0097] where, and are the mathematical expectation and variance of the statistic respectively, is a constant representing the multiple of the standard deviation. Conventionally, take , and set the rejection region to be outside . According to the properties of the normal distribution, the probability that the statistic falls within is 99.73%, and the probability of falling outside this range is only 0.27% (that is, the significance level ), the probability of falling outside one side of the range is only 0.135%, which is a small probability event. For a non-normal distribution, the probability of the statistic falling outside the range is also close to zero. According to the statistic of the process relative to the output of the process Depending on the different construction methods, quality control charts are mainly divided into three types, namely Shewhart control charts, Cumulative Sum (CUSUM) control charts, and Exponentially Weighted Moving Average (EWMA) control charts. The exponentially weighted moving average control chart adopted by the present invention is the EWMA control chart.

[0098] Time rate sequence The recursive form of the EWMA control chart statistic of

[0099] is shown in formula (2):

[0100] where is the smoothing coefficient, which is a constant and , Generally, take the mathematical expectation of all time rate sequences It can be expanded to represent the form of the exponentially weighted average of all time rate sequences, as shown in formula (3):

[0101] (3).

[0102] Assume that the time rate sequence is a sequence of random variables with the same distribution at different times. Then the estimated value of the mathematical expectation of the time rate sequence is , and the estimated value of the variance is

[0103] . It should be noted that in actual situations, the time rate sequence may contain outliers, and these outliers will cause biases in the estimation of the mean and will result in an overestimated maximum likelihood estimate value of the standard deviation. Therefore, if not adjusted, the effectiveness of the outlier detection method will decrease. To enhance the effectiveness of the test, the present invention adopts a robust method for estimating the mean and standard deviation to replace the conventional maximum likelihood estimation method. A robust mean is to calculate the average value after removing the largest and smallest 5% of the data at both ends after sorting the data; a robust method for estimating the standard deviation is shown in formula (4):

[0104] (4).

[0105] For The mathematical expectations and variances of both sides of the expansion formula (2) are calculated simultaneously, and after arrangement, the mathematical expectations and variances of the EWMA control chart statistics are shown in formula (5) respectively:

[0106] (5),

[0107] where n is the length of the time rate sequence. Finally, the upper / lower control limits (UCL / LCL) and the center line (CL) of the EWMA control chart are shown in formula (6) respectively:

[0108] (6),

[0109] where is a preset constant, indicating that the control limit is the mean plus / minus a multiple of the standard deviation. In this case, take 、 , and judge the upper / lower control limits of the EWMA control chart for the time rate sequence of dune movement pixel by pixel. If it exceeds the limit, it is determined that the rate is abnormal, that is, it is determined that the pixel has a rapid change during this time interval, and this point is marked in the time-series dune image.

[0110] The above dune movement rapid change detection method provided by the present invention has the following advantages compared with the prior art: First, the existing technology usually uses visual interpretation to manually mark and measure the dune movement distance in remote sensing images using geographic information processing software. The present invention can realize the automatic estimation of dune movement in remote sensing images by constructing the dune movement state through the optical flow method; Second, traditional dune movement measurement methods such as the five-point measurement method and the centroid measurement method are mostly used to construct the macroscopic movement of dunes. The present invention can estimate dune movement information from a microscopic perspective through the dense optical flow method; In addition, there are few current methods for detecting dune movement rapid changes. The present invention provides a simple and efficient detection method based on the optical flow method and the EWMA control chart, which can effectively detect the rapid changes in dune movement in remote sensing image sequences.

[0111] Figure 7 is a schematic structural diagram of a dune movement rapid change remote sensing detection system according to an embodiment of the present invention.

[0112] As Figure 7 shown, the above dune movement rapid change remote sensing detection system 700 includes a remote sensing image preprocessing module 710, a velocity vector field construction module 720, and a dune rapid change detection module 730.

[0113] The remote sensing image preprocessing module 710 is used to extract the dune edge from the time-series images of dune remote sensing observations by using deep learning methods, obtain a sequence of dune edge images, and preprocess the sequence of dune edge images to obtain a sequence of sand ridge line images; in one embodiment, the remote sensing image preprocessing module 710 can be used to perform the operation S210 described above, which will not be elaborated here.

[0114] The velocity vector field construction module 720 is used to construct a velocity vector field for the sand ridge line image data of two adjacent frames in the sequence of sand ridge line images by using the dense optical flow method, and obtain a spatio-temporal cube array of the sand ridge line moving speed field; in one embodiment, the velocity vector field construction module 720 can be used to perform the operation S220 described above, which will not be elaborated here.

[0115] The dune speed change detection module 730 is used to perform pixel-by-pixel time-series analysis on the spatio-temporal cube array of the sand ridge line moving speed field by using an exponentially weighted moving average control chart, obtain a judgment result of time-series rate abnormal points, and identify the position-time by superimposing the judgment result of the time-series rate abnormal points on the time-series images of dune remote sensing observations, so as to obtain the dune movement speed change detection result. In one embodiment, the dune speed change detection module 730 can be used to perform the operation S230 described above, which will not be elaborated here.

[0116] According to an embodiment of the present invention, any multiple of the remote sensing image preprocessing module 710, the velocity vector field construction module 720, and the dune speed change detection module 730 can be combined and implemented in one module, or any one of them can be split into multiple modules. Or, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present invention, at least one of the remote sensing image preprocessing module 710, the velocity vector field construction module 720, and the dune speed change detection module 730 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on chip, a system on substrate, a system on package, an application specific integrated circuit (ASIC), or can be implemented by any other reasonable way of integrating or packaging circuits and other hardware or firmware, or implemented in any one of the three implementation methods of software, hardware, and firmware or in any appropriate combination of several of them. Or, at least one of the remote sensing image preprocessing module 710, the velocity vector field construction module 720, and the dune speed change detection module 730 can be at least partially implemented as a computer program module, and when the computer program module is run, it can perform the corresponding functions.

[0117] The following is through specific experiments and in combination with the attached Figure 8 - attached Figure 10Verify the above methods and systems provided by the present invention.

[0118] Figure 8 It is a schematic diagram of the actual scenes of two deserts, Sand1 and Sand2, according to the embodiments of the present invention, and the corresponding schematic diagram of the extraction of sand ridge lines.

[0119] Figure 9 It is the input image and the corresponding output image of the rapid change detection of Sand1 according to the embodiments of the present invention.

[0120] Figure 10 It is a schematic diagram of the rapid change detection results of Sand1 during the years 2020 - 2021 according to the embodiments of the present invention.

[0121] Figure 11 It is a schematic diagram of the rapid change detection results of Sand1 during the years 2010 - 2014 according to the embodiments of the present invention.

[0122] Figure 12 It is the input image and the corresponding output image of the rapid change detection of Sand2 according to the embodiments of the present invention.

[0123] Figure 13 It is a schematic diagram of the rapid change detection results of Sand2 during the years 2010 - 2014 according to the embodiments of the present invention.

[0124] The data is from the experimental settings: The data is as Figure 8 shown, divided into two groups of data, Sand1 and Sand2. Each group of data consists of 6 actual scene images and sand ridge line extraction images of the same desert area, taken from left to right in 2010, 2014, 2016, 2020, 2021 to 2022. Since the movement of sand dunes is not a simple rigid body movement and involves a large number of fluid shape changes, the experiment uses the sand ridge line as the input and establishes an optical flow field with the movement changes of the sand ridge line in the image sequence for rapid change detection, including: the sudden appearance or disappearance of the sand ridge line, the change in the shape of the sand ridge line, and the rapid increase or rapid decrease in the movement speed of the sand dune. For the EWMA control chart, the parameter settings are as follows: standard deviation multiple , smoothing coefficient .

[0125] Dense optical flow method - EWMA detection results and analysis: Figure 9It shows the input and output of the rapid change detection of Sand1. The upper row is the input image sequence of sand ridge line extraction. It can be seen that the sand ridge line moves and changes in shape in the image sequence. The lower row is the output map of rapid change detection. The bottom map is the actual desert scene of this area. The green punctuation marks and arrows are the velocity field vectors established by the Farneback optical flow method, and the red part is the rapid change anomaly point. Since the Farneback optical flow method establishes the velocity field based on two adjacent frames of images and the velocity field is depicted on the bottom map of the second frame, the velocity field is not shown in the initial image, that is, the 2010 image.

[0126] Taking the rapid change detection results of Sand1 from 2020 to 2021 as an example for analysis, as Figure 10 shown, in the output result map, the yellow box is marked as the appearance of the sand ridge line. Comparing the sand ridge line input maps of 2020 and 2021, it can be clearly observed that the sand ridge line suddenly appears in the same area of the yellow box. Looking at the sand ridge line extraction maps from 2010 to 2020, the sand ridge line here shows a trend of gradually fading and disappearing, while it suddenly appears in the 2021 map, indicating a rapid change, which is consistent with the area determined as a rapid change by the algorithm; the blue box is marked as the rapid change of the sand ridge line movement. Comparing the sand ridge line input maps of 2020 and 2021, it can be observed that some of the sand ridge lines in the blue box area move and break towards the lower right direction. Looking at the sand ridge line images in previous years, the sand ridge line in this area has always shown continuous signs, while it breaks in the 2021 map, indicating a movement rapid change, which is also consistent with the area determined as a rapid change by the algorithm.

[0127] Taking the rapid change detection results of Sand1 from 2010 to 2014 as an example for analysis, as Figure 11 shown, in the output result map, the purple box is marked as the disappearance of the sand ridge line. Comparing the sand ridge line input maps of 2010 and 2014, it can be clearly observed that the sand ridge line suddenly disappears in the same area of the purple box. Looking at the sand ridge line extraction maps from other years, the sand ridge line has always existed in this area, while it suddenly disappears in 2014, indicating a rapid change, which is also consistent with the marked result of the algorithm.

[0128] The same test was carried out on the Sand2 sand dune data. Compared with the Sand1 data, there are more sand dunes in the Sand2 data, and this algorithm can still well construct the movement state of the sand dunes and can detect the rapid change points of the sand dune movement. The detection results are as Figure 12 shown, which proves the generalization performance of this algorithm.

[0129] Taking the rapid change detection results of Sand2 from 2010 to 2014 as an example for analysis, as Figure 13As shown, in the output result diagram, the yellow box indicates the appearance of sand ridges. Comparing the sand ridge input diagrams of 2010 and 2014, it can be clearly observed that the sand ridges suddenly appear in the same area of the yellow box. Looking at the sand ridge extraction diagrams from 2010 to 2022, the sand ridges here remain relatively stable in the following years. The sand ridges were missing here in 2010, but suddenly appeared in 2014, indicating a rapid change, which is consistent with the area determined by the algorithm as a rapidly changing area; the blue box indicates the rapid change in the movement of sand ridges. Comparing the sand ridge input diagrams of 2010 and 2014, it can be observed that some of the sand ridges in the blue box area move rapidly towards the lower right direction, which is consistent with the rapidly changing detection area estimated by the algorithm; the purple box indicates the disappearance of sand ridges. It can be clearly observed from the comparison of the images in 2010 and 2014 that the sand ridges disappear, which is also consistent with the algorithm estimation.

[0130] Figure 14 It is a block diagram of an electronic device suitable for implementing the method for remote sensing detection of rapid changes in dune movement according to an embodiment of the present invention.

[0131] As Figure 14 shown, the electronic device 1400 according to an embodiment of the present invention includes a processor 1401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1402 or a program loaded from a storage section 1408 into a random access memory (RAM) 1403. The processor 1401 may include, for example, a general microprocessor (such as a CPU), an instruction set processor, and / or a related chipset, and / or a dedicated microprocessor (such as an application specific integrated circuit (ASIC)), etc. The processor 1401 may also include on-board memory for caching purposes. The processor 1401 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0132] In the RAM 1403, various programs and data required for the operation of the electronic device 1400 are stored. The processor 1401, the ROM 1402, and the RAM 1403 are connected to each other through a bus 1404. The processor 1401 performs various operations of the method flow according to an embodiment of the present invention by executing the programs in the ROM 1402 and / or the RAM 1403. It should be noted that the program may also be stored in one or more memories other than the ROM 1402 and the RAM 1403. The processor 1401 may also perform various operations of the method flow according to an embodiment of the present invention by executing the programs stored in the one or more memories.

[0133] According to an embodiment of the present invention, the electronic device 1400 may further include an input / output (I / O) interface 1405, and the input / output (I / O) interface 1405 is also connected to the bus 1404. The electronic device 1400 may further include one or more of the following components connected to the input / output (I / O) interface 1405: an input portion 1406 including a keyboard, a mouse, etc.; an output portion 1407 including, for example, a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 1408 including a hard disk, etc.; and a communication portion 1409 including a network interface card such as a LAN card, a modem, etc. The communication portion 1409 performs communication processing via a network such as the Internet. The drive 1410 is also connected to the input / output (I / O) interface 1405 as needed. A removable medium 1411, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1410 as needed so that a computer program read therefrom can be installed into the storage portion 1408 as needed.

[0134] The present invention also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiment; or may exist separately without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present invention is implemented.

[0135] According to an embodiment of the present invention, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to an embodiment of the present invention, the computer-readable storage medium may include the above-described ROM 1402 and / or RAM 1403 and / or one or more memories other than ROM 1402 and RAM 1403.

[0136] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a portion of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions noted in the blocks may occur in a different order than that noted in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram or flowchart, and combinations of blocks in the block diagram or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or by a combination of dedicated hardware and computer instructions.

[0137] Those skilled in the art will appreciate that the features described in the various embodiments of the present invention can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in the present invention. In particular, without departing from the spirit and teachings of the present invention, the features described in the various embodiments of the present invention can be combined and / or combined in various ways. All such combinations and / or combinations fall within the scope of the present invention.

[0138] The embodiments of the present invention have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of the present invention. Although the embodiments have been described separately above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Without departing from the scope of the present invention, those skilled in the art can make various substitutions and modifications, and all such substitutions and modifications should fall within the scope of the present invention.

Claims

1. A remote sensing detection method for the variable speed of sand dune movement, characterized in that, The method includes: Using a deep learning method to extract the dune edge from the time-series remote sensing observation images of the dunes, obtaining a sequence of dune edge images, and preprocessing the sequence of dune edge images to obtain a sequence of ridge line images; Using the dense optical flow method to construct a velocity vector field for the ridge line image data of two adjacent frames in the sequence of ridge line images, obtaining a spatio-temporal cube array of the ridge line movement velocity field; Using an exponentially weighted moving average control chart to perform pixel-by-pixel time-series analysis on the spatio-temporal cube array of the ridge line movement velocity field, obtaining a judgment result of time-series rate abnormal points, and superimposing the judgment result of the time-series rate abnormal points on the time-series remote sensing observation images of the dunes for position-time marking, obtaining a dune movement speed change detection result.

2. The method according to claim 1, wherein Preprocessing the sequence of dune edge images to obtain a sequence of ridge line images includes: Using an erosion filtering method to perform noise filtering on the sequence of dune edge images, obtaining a denoised sequence of dune edge images; Based on the sequence of dune edge images, using a bilinear interpolation method to perform resampling on the denoised sequence of dune edge images, obtaining the sequence of ridge line images.

3. The method according to claim 1, wherein Using the dense optical flow method to construct a velocity vector field for the ridge line image data of two adjacent frames in the sequence of ridge line images, obtaining a spatio-temporal cube array of the ridge line movement velocity field includes: Using the Farneback dense optical flow model to calculate the horizontal displacement and vertical displacement of each pixel in the ridge line image data of two adjacent frames, obtaining a time series of the displacement vector field; Stitching the time series of the displacement vector fields of all the ridge line image data of two adjacent frames, obtaining a stitching result; Based on the stitching result, performing an operation on the displacement vector of each of the ridge line image data of two adjacent frames and the time interval of each of the ridge line image data of two adjacent frames, obtaining the spatio-temporal cube array of the ridge line movement velocity field.

4. The method according to claim 3, wherein The spatio-temporal cube array of the ridge line movement velocity field includes time-axis data, image horizontal-axis data, image vertical-axis data, pixel horizontal displacement data, pixel vertical displacement data, pixel horizontal velocity data, and pixel vertical velocity data, and the spatio-temporal cube array of the ridge line movement velocity field is used to characterize the velocity information of the pixels in the sequence of ridge line images.

5. The method according to claim 1, characterized in that, Using an exponentially weighted moving average control chart to perform pixel-by-pixel time-series analysis on the spatio-temporal cube array of the ridge line movement velocity field, obtaining a judgment result of time-series rate abnormal points includes: Extracting the displacement sequence and movement rate of each image coordinate on the ridge line from the spatio-temporal cube array of the ridge line movement velocity field, obtaining a spatio-temporal sequence of the ridge line movement rate; Using an exponentially weighted moving average control chart to perform discrimination of time-series rate abnormal points on the spatio-temporal sequence of the ridge line movement rate, obtaining the judgment result of the time-series rate abnormal points.

6. The method according to claim 5, characterized in that, Extracting the displacement sequence and movement rate of each image coordinate on the ridge line from the spatio-temporal cube array of the ridge line movement velocity field, obtaining a spatio-temporal sequence of the ridge line movement rate includes: Perform pixel-by-pixel labeling on the spatio-temporal cube array of the sand ridge line movement speed field at each moment to obtain the pixel labeling points of the sand ridge line at each moment. Add the horizontal axis velocity vector and the vertical axis velocity vector corresponding to the pixel labeling points of the sand ridge line at each moment, and then perform a modulus operation to obtain the spatio-temporal sequence of the sand ridge line movement rate.

7. The method according to claim 5, characterized in that, Use an exponentially weighted moving average control chart to discriminate the time-series rate anomaly points of the spatio-temporal sequence of the sand ridge line movement rate, and the obtained time-series rate anomaly point judgment results include: Construct a recurrence formula for the statistic of the exponentially weighted moving average control chart using the spatio-temporal sequence of the sand ridge line movement rate, and expand the recurrence formula of the statistic into the form of an exponentially weighted average of all time rate sequences to obtain the expansion formula of the statistic. Based on the maximum likelihood estimation method, construct an estimate of the standard deviation of the statistic, and use the estimate of the standard deviation of the statistic to operate on the expansion formula of the statistic to obtain the upper and lower control limits and the center line of the exponentially weighted moving average control chart. Use the upper and lower control limits and the center line of the exponentially weighted moving average control chart to perform pixel-by-pixel discrimination of the time-series rate anomaly points of the spatio-temporal sequence of the sand ridge line movement rate, and determine the pixels whose time-series rate exceeds the upper and lower control limits of the exponentially weighted moving average control chart as the time-series rate anomaly points.

8. A remote sensing detection system for variable sand dune movement speed, characterized in that, The system includes: A remote sensing image preprocessing module, which is used to extract the sand dune edge from the time-series images of sand dune remote sensing observations using a deep learning method to obtain a sequence of sand dune edge images, and preprocess the sequence of sand dune edge images to obtain a sequence of sand ridge line images. A velocity vector field construction module, which is used to construct a velocity vector field for the sand ridge line image data of two adjacent frames in the sequence of sand ridge line images using the dense optical flow method to obtain a spatio-temporal cube array of the sand ridge line movement speed field. A sand dune rapid change detection module, which is used to perform pixel-by-pixel time-series analysis on the spatio-temporal cube array of the sand ridge line movement speed field using an exponentially weighted moving average control chart to obtain the time-series rate anomaly point judgment results, and perform position-time identification by superimposing the judgment results of the time-series rate anomaly points on the time-series images of sand dune remote sensing observations to obtain the sand dune movement rapid change detection results.

9. An electronic device, including: One or more processors; A memory for storing one or more computer programs, Characterized in that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instruction is executed by the processor, it implements the steps of the method according to any one of claims 1 to 7.

Images