Method, device, computer equipment and storage medium for monitoring flooding range

Abstract

The present application relates to a method, device, computer equipment, storage medium and computer program product for monitoring flooding range. The method includes: for each synthetic aperture radar satellite image of the target monitoring area within the monitoring period, based on the synthetic aperture radar satellite image and a preset generalized Gaussian mixture model, a probability model of the backscatter coefficient of the synthetic aperture radar satellite image is established; based on the probability model and the expectation maximization E‑M algorithm, the classification threshold corresponding to the synthetic aperture radar satellite image is solved; based on the comparison result of the backscatter coefficient observed for each pixel in the synthetic aperture radar satellite image and the classification threshold, the target category corresponding to each pixel is determined; according to the target category corresponding to each pixel, the first real-time flooded area corresponding to the synthetic aperture radar satellite image is calculated. The use of this method can expand the observation coverage of flooding range monitoring and evaluate the evolution of floods in flood storage areas and the changes in flooding range over time.

Description

Technical Field

[0001] The present application relates to the field of satellite remote sensing technology, and in particular to a method, device, computer equipment, storage medium and computer program product for monitoring flooding range. Background Art

[0002] Flood storage and detention areas refer to low-lying areas or lakes that include levees, flood diversion facilities, and flood discharge facilities, which are used to temporarily store flood water or discharge flood peaks. Under the influence of human activities such as climate change and urbanization, heavy rainfall and floods in the region are becoming more frequent, and flood control safety is facing new and severe challenges. Flood storage and detention areas are an important measure to properly handle the relationship between the local and the overall situation and minimize the overall losses caused by flood disasters by timely storing excess flood water and reducing flood peaks. It is crucial to know the evolution of floods in flood storage and detention areas in real time and make correct decisions to protect people's lives and property.

[0003] The relevant methods for obtaining the evolution of floods in flood storage areas are mainly based on monitoring methods. The relevant monitoring methods for the inundation range are mainly based on field measurement methods. Field measurement methods are often based on field photography by drones, with ultra-high spatial resolution and flexible observation time, and can quickly carry out monitoring for sudden floods.

[0004] However, the relevant inundation range monitoring methods have a small observation coverage, and it is usually difficult to ensure consistent and continuous observation of a larger area of ​​flood storage and detention areas over a multi-day period, making it difficult to assess the evolution of floods and changes in inundation range over time in flood storage and detention areas. Summary of the invention

[0005] Based on this, it is necessary to provide a monitoring method, device, computer equipment, computer-readable storage medium and computer program product for the flooding range that can expand the observation coverage in response to the above-mentioned technical problems.

[0006] In a first aspect, the present application provides a method for monitoring a flooded area, comprising:

[0007] For each synthetic aperture radar satellite image of the target monitoring area within the monitoring period, a probability model of the backscattering coefficient of the synthetic aperture radar satellite image is established based on the synthetic aperture radar satellite image and a preset generalized Gaussian mixture model;

[0008] Based on the probability model and the expectation-maximization EM algorithm, solving the classification threshold corresponding to the synthetic aperture radar satellite image;

[0009] Based on the comparison result of the backscatter coefficient observed for each pixel in the synthetic aperture radar satellite image and the classification threshold, determining the target category corresponding to each pixel; the target category is one of a water body and a non-water body;

[0010] Calculating a first real-time flooded area corresponding to the synthetic aperture radar satellite image according to the target category corresponding to each pixel;

[0011] The flooded area monitoring result of the target monitoring area is determined according to the first real-time flooded area corresponding to each of the synthetic aperture radar satellite images.

[0012] In one embodiment, solving the classification threshold corresponding to the synthetic aperture radar satellite image based on the probability model and the expectation maximization EM algorithm includes:

[0013] Using the expectation maximization (EM) algorithm, the probability model is calibrated to obtain calibrated parameters corresponding to the synthetic aperture radar satellite image;

[0014] Based on the calibrated parameters and a preset numerical algorithm, a classification threshold corresponding to the synthetic aperture radar satellite image is solved.

[0015] In one embodiment, the method of using the expectation maximization (EM) algorithm to calibrate the parameters of the probability model to obtain the calibrated parameters corresponding to the synthetic aperture radar satellite image includes:

[0016] For the first synthetic aperture radar satellite image, multiple initial values ​​of the parameters to be calibrated of the probability model are tested, and a first optimal initial value set is selected; the optimal initial value set includes a first optimal initial value of a scaling parameter, a first optimal initial value of a shape parameter, a first optimal initial value of a mean, and a first optimal initial value of a category pixel ratio;

[0017] Using the expectation maximization (EM) algorithm and the optimal initial value set, the probability model is calibrated to obtain calibrated parameters corresponding to the first synthetic aperture radar satellite image;

[0018] For subsequent synthetic aperture radar satellite images within the monitoring period, the first optimal initial value of the scaling parameter, the first optimal initial value of the shape parameter, and the first optimal initial value of the category pixel ratio are used as the second optimal initial value of the scaling parameter, the second optimal initial value of the shape parameter, and the second optimal initial value of the category pixel ratio, respectively, and multiple initial values ​​of the mean of the probability model are tested to select the second optimal initial value of the mean;

[0019] The probability model is calibrated using an expectation-maximization EM algorithm, the second optimal initial value of the scaling parameter, the second optimal initial value of the shape parameter, the second optimal initial value of the mean, and the second optimal initial value of the category pixel ratio to obtain calibrated parameters corresponding to the subsequent synthetic aperture radar satellite image.

[0020] In one embodiment, determining the flooded area monitoring result of the target monitoring area according to the first real-time flooded area corresponding to each synthetic aperture radar satellite image includes:

[0021] Acquire an optical satellite image of the target monitoring area within the monitoring period; a first image time corresponding to the synthetic aperture radar satellite image is different from a second image time corresponding to the optical satellite image;

[0022] For each optical satellite image, determining a second real-time flooded area corresponding to the optical satellite image according to the optical satellite image and a preset random forest algorithm;

[0023] The first real-time flooded area corresponding to each synthetic aperture radar satellite image and the second real-time flooded area corresponding to each optical satellite image constitute the flooded area monitoring result of the target monitoring area.

[0024] In one embodiment, determining the second real-time flooded area corresponding to the optical satellite image according to the optical satellite image and a preset random forest algorithm includes:

[0025] Determining training sample points of the optical satellite image in the optical satellite image according to the geographic feature information of the target monitoring area;

[0026] Based on the training sample points and a preset random forest algorithm, a random forest model corresponding to the optical satellite image is established;

[0027] A second real-time flooded area corresponding to the optical satellite image is determined according to the random forest model and the optical satellite image.

[0028] In one embodiment, determining the training sample points of the optical satellite image in the optical satellite image according to the geographic feature information of the target monitoring area includes:

[0029] If the geographic feature information of the target monitoring area indicates the existence of small water bodies and / or field ridges, and the spatial resolution of the optical satellite image is lower than a preset resolution threshold, then according to the spatial position information of the center points of the small water bodies and / or field ridges in the synthetic aperture radar satellite images and / or other optical satellite images adjacent to the optical satellite image and having a spatial resolution higher than the preset resolution threshold, determine the center points of the small water bodies and / or field ridges in the optical satellite image, and generate training sample points including the center points of the small water bodies and / or field ridges in the optical satellite image;

[0030] If the geographic feature information of the target monitoring area indicates that there is a water flow velocity greater than a preset flow velocity threshold, and the water turbidity of the optical satellite image is greater than a preset water turbidity threshold, then according to the spatial position information of each point in the synthetic aperture radar satellite image and / or other optical satellite image whose water turbidity adjacent to the optical satellite image is less than or equal to the preset water turbidity threshold, determine the training sample points of the optical satellite image in the optical satellite image;

[0031] If the geographic feature information of the target monitoring area indicates that the proportion of the flooded area is less than a preset proportion threshold, the selection range of the training sample points of the optical satellite image is determined according to the cloud shadow shape and the building shadows of the synthetic aperture radar satellite images and / or other optical satellite images at different times on adjacent dates, and the training sample points of the optical satellite image are determined within the selection range.

[0032] In a second aspect, the present application also provides a flood range monitoring device, comprising:

[0033] A modeling module is used to establish a probability model of the backscattering coefficient of each synthetic aperture radar satellite image of the target monitoring area within the monitoring period based on the synthetic aperture radar satellite image and a preset generalized Gaussian mixture model;

[0034] A solution module, used for solving the classification threshold corresponding to the synthetic aperture radar satellite image based on the probability model and the expectation maximization EM algorithm;

[0035] A first determination module is used to determine the target category corresponding to each pixel based on the comparison result of the backscatter coefficient observed by each pixel in the synthetic aperture radar satellite image and the classification threshold; the target category is one of water body and non-water body;

[0036] A calculation module, used for calculating a first real-time flooded area corresponding to the synthetic aperture radar satellite image according to the target category corresponding to each pixel;

[0037] The second determination module is used to determine the flooded area monitoring result of the target monitoring area according to the first real-time flooded area corresponding to each synthetic aperture radar satellite image.

[0038] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps described in the first aspect when executing the computer program.

[0039] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps described in the first aspect above.

[0040] In a fifth aspect, the present application further provides a computer program product, comprising a computer program, which implements the steps described in the first aspect when executed by a processor.

[0041] The above-mentioned monitoring method, device, computer equipment, storage medium and computer program product for the flooding range, for each synthetic aperture radar satellite image of the target monitoring area within the monitoring period, establish a probability model of the backscattering coefficient of the synthetic aperture radar satellite image based on the synthetic aperture radar satellite image and a preset generalized Gaussian mixture model; solve the classification threshold corresponding to the synthetic aperture radar satellite image based on the probability model and the expectation-maximization EM algorithm; determine the target category corresponding to each pixel in the synthetic aperture radar satellite image based on the comparison result of the backscattering coefficient observed for each pixel in the synthetic aperture radar satellite image with the classification threshold; the target category is one of water body and non-water body; according to the target category corresponding to each pixel, calculate the first real-time flooding area corresponding to the synthetic aperture radar satellite image; according to the first real-time flooding area corresponding to each synthetic aperture radar satellite image, determine the flooding area monitoring result of the target monitoring area. In this way, synthetic aperture radar satellite images of the target monitoring area are obtained, and a generalized Gaussian mixture model of the backscatter coefficient is established for each image. Then, based on the model and the EM algorithm, the classification threshold corresponding to each model for distinguishing whether a pixel is a water body is solved. By comparing the backscatter coefficient of the pixel and the classification threshold, it is determined whether each pixel is a water body, and the real-time flooded area corresponding to each image is calculated to obtain the flooded area monitoring results. Compared with drones, satellite remote sensing has expanded the observation coverage of flood range monitoring, which can ensure consistent and continuous observation of a larger range of flood storage and detention areas over a multi-day period, and can evaluate the flood evolution and flood range changes over time in flood storage and detention areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1 A schematic diagram of a flow chart of a method for monitoring a flooded area in one embodiment;

[0044] Figure 2 A schematic diagram of a flow chart of a step of solving a classification threshold value corresponding to the synthetic aperture radar satellite image in an embodiment;

[0045] Figure 3 A schematic diagram of a flow chart of a step of performing parameter calibration on a probability model in one embodiment;

[0046] Figure 4 A schematic flow chart of a step of determining a flooded area monitoring result of a target monitoring area in an embodiment;

[0047] Figure 5 A schematic diagram of a flow chart of a step of determining a second real-time flooded area corresponding to an optical satellite image in an embodiment;

[0048] Figure 6 A schematic diagram of a flow chart of a step of determining training sample points of an optical satellite image in one embodiment;

[0049] Figure 7 A schematic diagram of monitoring results of the submerged area of ​​flood storage area B in an embodiment;

[0050] Figure 8 is a structural block diagram of a monitoring device for flooding range in one embodiment;

[0051] Fig. 9 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0053] In one embodiment, Figure 1As shown, a method for monitoring the flooding range is provided. This embodiment uses the method applied to a terminal as an example for illustration. It is understandable that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. Among them, the terminal can be but is not limited to various personal computers, laptops, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server can be implemented as an independent server or a server cluster composed of multiple servers. In this embodiment, the method includes the following steps:

[0054] Step 101, for each synthetic aperture radar satellite image of the target monitoring area within the monitoring period, a probability model of the backscattering coefficient of the synthetic aperture radar satellite image is established based on the synthetic aperture radar satellite image and a preset generalized Gaussian mixture model.

[0055] In an embodiment of the present application, the target monitoring area is an area to be monitored for flooded area, which may be a flood storage area or other area that has encountered floods. The monitoring time period is a time period for monitoring the evolution of the flooded area. The synthetic aperture radar satellite image is an image of a synthetic aperture radar satellite, which may be an image of one or more synthetic aperture radar satellites. The synthetic aperture radar (SAR) satellite may be a domestic satellite or a foreign satellite. For example, the synthetic aperture radar satellite image may be a Gaofen-3 satellite image and a Sentinel-1 satellite image. The synthetic aperture radar satellite image may be a satellite image of a daily scale or a satellite image taken multiple times a day. In the case where the synthetic aperture radar satellite image is an image of multiple synthetic aperture radar satellites, the first image time corresponding to different types of synthetic aperture radar satellite images is different. In one example, the first image time corresponding to one synthetic aperture radar satellite image is earlier or later than the first image time corresponding to another synthetic aperture radar satellite image. In one example, the first image time corresponding to one synthetic aperture radar satellite image may be located between the first image times corresponding to two adjacent synthetic aperture radar satellite images of another synthetic aperture radar satellite image. For example, the first image times corresponding to all satellite images in one synthetic aperture radar satellite image may be located between the first image times corresponding to a group of two adjacent synthetic aperture radar satellite images in another synthetic aperture radar satellite image. For example, the first image time corresponding to each optical satellite image in one synthetic aperture radar satellite image may be located between the first image times corresponding to two different adjacent synthetic aperture radar satellite images in another synthetic aperture radar satellite image.

[0056] In one embodiment, within the target monitoring area A, the Gaofen-3 satellite image provides horizontal-horizontal HH polarization and horizontal-vertical HV polarization backscatter coefficients, and the Sentinel-1 satellite image provides vertical-vertical VV polarization and vertical-horizontal VH polarization backscatter coefficients. Among them, the non-co-directional HV and VH polarization backscatter coefficients are weak in distinguishing water bodies from land. Therefore, this embodiment uses the HH polarization backscatter coefficient of the Gaofen-3 satellite image and the VV polarization backscatter coefficient of the Sentinel-1 satellite image, and assumes that they obey a generalized Gaussian mixture distribution respectively. The generalized Gaussian mixture model assumes that the population can be divided into k categories (k=2), each of which obeys a generalized Gaussian distribution, which can be expressed as:

[0057]

[0058]

[0059] Where x is the backscatter coefficient observed for a certain pixel (HH polarization for Gaofen-3 satellite images and VV polarization for Sentinel-1 satellite images), which is a scalar; k is the category to which the pixel belongs. If k is 0, it is a non-water pixel, and k is 1, it is a water pixel; is the conditional probability that the observed backscatter coefficient is equal to x when the pixel belongs to the kth class; θ k =[α k ,β k ,μ k ] is the parameter of the generalized Gaussian distribution; is the gamma function; p(x) is the probability that the VV or HH polarization backscatter coefficient observed for a certain pixel is x; p k is the proportion of the k-th class of pixels in the population, that is, the probability of a randomly selected pixel belonging to the k-th class.

[0060] Step 102: Based on the probability model and the expectation-maximization (EM) algorithm, a classification threshold corresponding to the synthetic aperture radar satellite image is solved.

[0061] In the embodiment of the present application, the expectation-maximization (EM) algorithm is used to calibrate the parameters of the probability model of the backscatter coefficient of the synthetic aperture radar satellite image. The classification threshold is the threshold of the backscatter coefficient, which is used to distinguish whether the pixels corresponding to different backscatter coefficients are water bodies.

[0062] Step 103, based on the comparison result of the backscatter coefficient observed for each pixel in the synthetic aperture radar satellite image and the classification threshold, determine the target category corresponding to each pixel.

[0063] The target category is one of water body and non-water body.

[0064] In the embodiment of the present application, for each pixel in the synthetic aperture radar satellite image, the terminal compares the backscatter coefficient observed by the pixel with the classification threshold. If the backscatter coefficient observed by the pixel is less than the classification threshold, the terminal determines that the target category corresponding to the pixel is a water body. If the backscatter coefficient observed by the pixel is greater than or equal to the classification threshold, the terminal determines that the target category corresponding to the pixel is a non-water body.

[0065] In one embodiment, the terminal determines the target category corresponding to each pixel based on the comparison result of the backscatter coefficient observed by each pixel in the synthetic aperture radar satellite image and the classification threshold, which can be expressed as:

[0066]

[0067]

[0068] in, is the classification result of pixel i, j of SAR satellite, x opt is the classification threshold of the backscatter coefficient, x i,j is the backscatter coefficient observed for a certain pixel.

[0069] Step 104: Calculate a first real-time flooded area corresponding to the synthetic aperture radar satellite image according to the target category corresponding to each pixel.

[0070] In the embodiment of the present application, the terminal uses the pixel whose target category is water body as the target pixel. Then, the terminal calculates the first real-time flooded area corresponding to the synthetic aperture radar satellite image according to the target pixel in the synthetic aperture radar satellite image. The first real-time flooded area is the real-time flooded area of ​​the target monitoring area obtained based on the SAR satellite image.

[0071] Step 105 , determining the flooded area monitoring result of the target monitoring area according to the first real-time flooded area corresponding to each synthetic aperture radar satellite image.

[0072] In the embodiment of the present application, the flooded area monitoring result includes the real-time flooded areas corresponding to multiple times of the target monitoring area within the monitoring period. The terminal generates the flooded area monitoring result of the target monitoring area including the first real-time flooded areas corresponding to each synthetic aperture radar satellite image.

[0073] In one example, the terminal uses the first real-time flooded area corresponding to each synthetic aperture radar satellite image to form a flooded area monitoring result of the target monitoring area.

[0074] In another example, the terminal obtains an optical satellite image of the target monitoring area within the monitoring period. Then, for each optical satellite image, the terminal determines the second real-time flooded area corresponding to the optical satellite image based on the optical satellite image. Then, the terminal generates a flooded area monitoring result of the target monitoring area including the first real-time flooded area corresponding to each synthetic aperture radar satellite image and the second real-time flooded area corresponding to each optical satellite image.

[0075] The above-mentioned method for monitoring the flooded range obtains synthetic aperture radar satellite images of the target monitoring area, establishes a generalized Gaussian mixture model of the backscatter coefficient for each image, and then solves the classification threshold corresponding to each model for distinguishing whether the pixel is a water body based on the model and the EM algorithm. By comparing the backscatter coefficient of the pixel and the classification threshold, it is determined whether each pixel is a water body, and the real-time flooded area corresponding to each image is calculated to obtain the flooded area monitoring result. The use of satellite remote sensing has expanded the observation coverage of flooded range monitoring compared to drones, and can ensure consistent and continuous observation of a larger range of flood storage and detention areas over a multi-day period, and can evaluate the flood evolution and flooded range changes over time in the flood storage and detention areas. In addition, this method establishes a generalized Gaussian mixture model and solves the generalized Gaussian mixture model to obtain the classification threshold. The assumption is more in line with the actual situation of the evolution of the flooded range of the flood storage area. Compared with the simple Gaussian distribution and the Otsu method (OTSU) commonly used for satellite image processing, the selection of the classification threshold is more stable and accurate, which can ensure the consistency of the time series of the flooded area obtained based on the images obtained at different times by the same satellite, and can also ensure the consistency of the time series of the flooded area obtained by the images obtained by different satellites at the same time. At the same time, the monitoring accuracy of the flooded range is higher and the uncertainty is smaller. Moreover, this method can realize the monitoring of the flooded range through domestic satellites, which can improve the autonomous controllability of data and thus improve the temporal resolution. Compared with satellite monitoring at the monthly and annual scales, this method can realize the monitoring of the flooded range at the daily scale and multiple times a day, and realize quasi-real-time, large-scale, high-frequency and high-efficiency monitoring of the flooded range.

[0076] In one embodiment, Figure 2 As shown, based on the probability model and the expectation maximization EM algorithm, the specific process of solving the classification threshold corresponding to the synthetic aperture radar satellite image includes the following steps:

[0077] Step 201, using the expectation maximization (EM) algorithm, calibrates the parameters of the probability model to obtain the calibrated parameters corresponding to the synthetic aperture radar satellite image.

[0078] In the embodiment of the present application, the parameters to be calibrated of the probability model are scaling parameter, shape parameter, mean value and category pixel ratio. The analog pixel ratio is the proportion of the category pixels in the total.

[0079] In one example, the terminal inputs the preset initial values ​​corresponding to the parameters to be calibrated of the probability model into the EM algorithm model, calibrates the parameters of the probability model, and obtains the calibrated parameters corresponding to the synthetic aperture radar satellite image.

[0080] In one example, the terminal tests multiple initial values ​​of the parameters to be calibrated of the probability model, and selects the optimal initial value of the scaling parameter, the optimal initial value of the shape parameter, the optimal initial value of the mean, and the optimal initial value of the category pixel ratio. Then, the terminal uses the expectation maximization EM algorithm, the optimal initial value of the scaling parameter, the optimal initial value of the shape parameter, the optimal initial value of the mean, and the optimal initial value of the category pixel ratio to calibrate the parameters of the probability model and obtain the calibrated parameters corresponding to the synthetic aperture radar satellite image.

[0081] Step 202, based on the calibrated parameters and a preset numerical algorithm, solve the classification threshold corresponding to the synthetic aperture radar satellite image.

[0082] In an embodiment of the present application, the terminal can solve the equation shown in the target formula to solve the classification threshold corresponding to the synthetic aperture radar satellite image. The target formula can be expressed as:

[0083]

[0084] Since p in the target formula k The term contains a gamma function, which cannot be solved analytically and can only be solved by numerical algorithms. This embodiment uses the Newton-Raphson algorithm (also known as the Newton iteration method). k is the proportion of the k-th class of pixels in the population, that is, the probability of a randomly selected pixel belonging to the k-th class; θ k =[α k ,β k ,μ k ] is the parameter of the generalized Gaussian distribution, α k is the scaling parameter, β k is the shape parameter, μ k is the mean.

[0085] The above-mentioned monitoring method of the flooded area adopts the expectation maximization EM algorithm to calibrate the parameters of the probability model to obtain the calibrated parameters corresponding to the synthetic aperture radar satellite image; based on the calibrated parameters and the preset numerical algorithm, the classification threshold corresponding to the synthetic aperture radar satellite image is solved. In this way, by modeling each satellite image in turn, calibrating the parameters and solving the classification threshold, this method is more in line with the actual situation that the classification threshold is not fixed but changes with time, which can further improve the stability of the classification threshold and further improve the consistency and accuracy of the time series of flooded area monitoring.

[0086] In one embodiment, Figure 3 As shown in FIG. 1 , the expectation maximization EM algorithm is used to calibrate the parameters of the probability model, and the specific process of obtaining the calibrated parameters corresponding to the synthetic aperture radar satellite image includes the following steps:

[0087] Step 301 : for a first synthetic aperture radar satellite image, test multiple initial values ​​of parameters to be calibrated of a probability model, and select a first optimal initial value set.

[0088] The optimal initial value set includes the first optimal initial value of the scaling parameter, the first optimal initial value of the shape parameter, the first optimal initial value of the mean, and the first optimal initial value of the category pixel ratio.

[0089] Step 302, using the expectation maximization EM algorithm and the optimal initial value set, calibrate the parameters of the probability model to obtain the calibrated parameters corresponding to the first synthetic aperture radar satellite image.

[0090] Step 303, for subsequent synthetic aperture radar satellite images within the monitoring period, the first optimal initial value of the scaling parameter, the first optimal initial value of the shape parameter, and the first optimal initial value of the category pixel ratio are used as the second optimal initial value of the scaling parameter, the second optimal initial value of the shape parameter, and the second optimal initial value of the category pixel ratio, respectively, and multiple initial values ​​of the mean of the probability model are tested to select the second optimal initial value of the mean.

[0091] Step 304, using the expectation-maximization EM algorithm, the second optimal initial value of the scaling parameter, the second optimal initial value of the shape parameter, the second optimal initial value of the mean, and the second optimal initial value of the category pixel ratio, the probability model is calibrated to obtain the calibrated parameters corresponding to the subsequent synthetic aperture radar satellite image.

[0092] In an embodiment of the present application, the first synthetic aperture radar satellite image is the synthetic aperture radar satellite image with the earliest image time among the synthetic aperture radar satellite images. Subsequent synthetic aperture radar satellite images are other synthetic aperture radar satellite images except the first synthetic aperture radar satellite image among the synthetic aperture radar satellite images. The image time may be the shooting time of the satellite image. It is understandable that any method of selecting a first optimal initial value set for any multiple initial values ​​of the parameters to be calibrated that can test the probability model can be used for step 301, any method of selecting a second optimal initial value of the mean that can test the probability model can be used for step 303, and any method of using the expectation maximization EM algorithm, the optimal initial value of the scaling parameter, the optimal initial value of the shape parameter, the optimal initial value of the mean, and the optimal initial value of the proportion of class pixels can be used for steps 302 and 304, and this application does not limit it.

[0093] The above-mentioned method for monitoring the flood range, the flood storage area flood range monitoring task targeted by this application, has extremely high requirements for timeliness and consistency, and the timeliness cannot be guaranteed by acquiring all the SAR image data within the study period at one time. The EM algorithm fitting results are sensitive to the selection of initial parameter values ​​to a certain extent, and the satellite images have quality fluctuations, resulting in separate modeling of each SAR image, initial value selection, and parameter calibration. The stability of the classification threshold and the consistency of the extraction results at different times cannot be guaranteed. This method proposes the above-mentioned updated modeling and parameter calibration process applicable to the timeliness and consistency requirements of flood storage and detention area inundation range monitoring. The first SAR image of the flood storage and detention area inundation range monitoring mission period is selected, all pixels within the flood storage and detention area boundary are sampled, a generalized Gaussian mixture model is established, and different parameter initial values ​​are tested. The optimal initial value set is selected, and all parameters are calibrated. For subsequent images of the same flood storage and detention area, it is reasonably assumed that the scaling parameters and shape parameters related to the properties of the category to which the pixels belong will not change. Although the proportion of pixels of a certain category in the total will change, it will not exceed the robustness range of the EM algorithm. For subsequent period images, when selecting the initial values ​​of the EM algorithm, the scaling parameters, shape parameters, and the proportion of category pixels are fixed to be the same as the optimal initial values ​​of the first image, and only the mean is adjusted, which is more in line with the actual situation, can further improve the stability of the classification threshold, and further improve the consistency of the time series of inundation range monitoring.

[0094] In one embodiment, Figure 4 As shown, the specific process of determining the monitoring result of the flooded area of ​​the target monitoring area according to the first real-time flooded area corresponding to each synthetic aperture radar satellite image includes the following steps:

[0095] Step 401, obtaining an optical satellite image of a target monitoring area within a monitoring period.

[0096] The first image time corresponding to the synthetic aperture radar satellite image is different from the second image time corresponding to the optical satellite image.

[0097] In the embodiment of the present application, the optical satellite image is an image of an optical satellite, which may be an image of one or more optical satellites. The optical satellite may be a domestic satellite or a foreign satellite. The optical satellite image may be a satellite image of a daily scale or a satellite image taken multiple times a day. The first image time and the second image time are both the shooting time of the satellite image. For example, the optical satellite image may be a Gaofen-1 satellite image, a Gaofen-6 satellite image, and an environmental disaster reduction satellite image. Gaofen-1 includes the 01 satellite launched in 2013 and the 02, 03, and 04 satellites launched simultaneously in 2018, equipped with two 2-meter resolution panchromatic / 8-meter resolution multispectral cameras (high-resolution cameras PMS) and four 16-meter resolution multispectral cameras (wide-width cameras WFV). The terminal may use multispectral images and panchromatic images to fuse 2-meter spatial resolution images and 16-meter resolution multispectral camera images, respectively, according to the actual available image conditions. Gaofen-6 was also launched in 2018, and this embodiment applies its 16-meter resolution multispectral camera image. Environmental Disaster Reduction No. 2A and No. 2B satellites were launched simultaneously in 2020, and this embodiment uses its 16-meter resolution charge-coupled device (CCD) camera images. In the case where the optical satellite images are images of multiple optical satellites, the second image times corresponding to different types of optical satellite images are different. In one example, the second image time corresponding to one optical satellite image is earlier or later than the second image time corresponding to another optical satellite image. In one example, the second image time corresponding to one optical satellite image can be located between the second image times corresponding to two adjacent optical satellite images of another optical satellite image. For example, the second image times corresponding to all satellite images in one optical satellite image can be located between the second image times corresponding to a group of two adjacent optical satellite images in another optical satellite image. For example, the second image time corresponding to each optical satellite image in one optical satellite image can be located between the second image times corresponding to two different adjacent optical satellite images in another optical satellite image.

[0098] In one example, the first image time corresponding to each synthetic aperture radar satellite image is earlier or later than the second image time corresponding to each optical satellite image.

[0099] In one example, a first image time corresponding to a synthetic aperture radar satellite image and a second image time corresponding to an optical satellite image may be on the same day.

[0100] In one example, the second image time corresponding to each optical satellite image may be between the first image times corresponding to two adjacent synthetic aperture radar satellite images. For example, the second image time corresponding to all optical satellite images may be between the first image times corresponding to a group of two adjacent synthetic aperture radar satellite images. For example, the second image time corresponding to each optical satellite image may be between the first image times corresponding to two different adjacent synthetic aperture radar satellite images.

[0101] Step 402 : for each optical satellite image, determine a second real-time flooded area corresponding to the optical satellite image according to the optical satellite image and a preset random forest algorithm.

[0102] In an embodiment of the present application, for each optical satellite image, the terminal establishes a random forest model corresponding to the optical satellite image based on the optical satellite image and a preset random forest algorithm. Then, the terminal inputs the optical satellite image into the random forest model corresponding to the optical satellite image to obtain the target type of each pixel in the optical satellite image. Then, the terminal calculates the second real-time flooded area corresponding to the optical satellite image based on the target type of each pixel in the optical satellite image. The second real-time flooded area is the real-time flooded area of ​​the target monitoring area obtained based on the optical satellite image.

[0103] Step 403: The first real-time flooded area corresponding to each synthetic aperture radar satellite image and the second real-time flooded area corresponding to each optical satellite image are combined to form a flooded area monitoring result of the target monitoring area.

[0104] The above-mentioned method for monitoring the flooded range obtains optical satellite images with different image times corresponding to the synthetic aperture radar satellite images, and also determines the real-time flooded area of ​​the target monitoring area based on the optical satellite images. The real-time flooded area obtained based on the two types of satellite images is integrated, and the time resolution and accuracy of flooded range monitoring can be further improved by assisting the SAR satellite through the optical satellite. It can also avoid problems such as fewer effective observations due to reliance on a single data source, thereby improving the effectiveness of flooded range monitoring.

[0105] In one embodiment, Figure 5 As shown, according to the optical satellite image and the preset random forest algorithm, the specific process of determining the second real-time flooded area corresponding to the optical satellite image includes the following steps:

[0106] Step 501, determining training sample points of the optical satellite image in the optical satellite image according to the geographical feature information of the target monitoring area.

[0107] In the embodiment of the present application, the terminal determines the sample point type of the optical satellite image according to the geographical feature information of the target monitoring area, and then selects the training sample points of the optical satellite image in the optical satellite image according to the sample point type of the optical satellite image.

[0108] Step 502: Establish a random forest model corresponding to the optical satellite image based on the training sample points and a preset random forest algorithm.

[0109] In the embodiment of the present application, it can be understood that any algorithm that can establish a random forest model based on training sample points can be used in step 502, and the present application does not limit it.

[0110] Step 503: Determine a second real-time flooded area corresponding to the optical satellite image based on the random forest model and the optical satellite image.

[0111] In an embodiment of the present application, geographic feature information is used to characterize the geographic features of the target monitoring area. The training sample points are points on the optical satellite image currently acquired by the terminal. The terminal constructs a random forest model corresponding to the optical satellite image based on the training sample points on the optical satellite image. The random forest model corresponding to the optical satellite image is used to identify whether the category to which each pixel on the optical satellite image belongs is a water body.

[0112] The terminal inputs the optical satellite image into the random forest model corresponding to the optical satellite image to obtain the target type of each pixel in the optical satellite image. Then, the terminal calculates the second real-time flooded area corresponding to the optical satellite image according to the target type of each pixel in the optical satellite image.

[0113] The above-mentioned method for monitoring the flooded range determines the training sample points of the optical satellite image in the optical satellite image according to the geographical feature information of the target monitoring area; establishes the random forest model corresponding to the optical satellite image based on the training sample points and the preset random forest algorithm; and determines the second real-time flooded area corresponding to the optical satellite image according to the random forest model and the optical satellite image. In this way, by establishing a respective random forest model for each optical satellite image, it is used to distinguish whether each pixel in its satellite image is a water body, and the calculation of the second real-time flooded area is realized, which can avoid the systematic differences in the results after the radiation calibration of different images, and the consistency differences caused by the lack of radiation calibration parameters of some sensors, and can improve the accuracy of the second real-time flooded area and the consistency of the monitoring of the flooded range. In addition, this method considers the influence of different geographical features of different target monitoring areas on water body identification, selects training sample points, and is more in line with the actual situation, which can further improve the accuracy of the second real-time flooded area and the accuracy of the monitoring of the flooded range.

[0114] In one embodiment, Figure 6 As shown, according to the geographic feature information of the target monitoring area, the specific process of determining the training sample points of the optical satellite image in the optical satellite image includes the following steps:

[0115] Step 601: If the geographic feature information of the target monitoring area indicates the existence of small water bodies and / or field ridges, and the spatial resolution of the optical satellite image is lower than a preset resolution threshold, the center point of the small water body and / or field ridge in the optical satellite image is determined based on the spatial position information of the center point of the small water body and / or field ridge in the synthetic aperture radar satellite image and / or other optical satellite images adjacent to the optical satellite image whose spatial resolution is higher than the preset resolution threshold, and training sample points containing the center point of the small water body and / or field ridge in the optical satellite image are generated.

[0116] Step 602: If the geographic feature information of the target monitoring area indicates that there is a water flow velocity greater than a preset flow velocity threshold and the water turbidity of the optical satellite image is greater than a preset water turbidity threshold, then based on the spatial position information of each point in the synthetic aperture radar satellite image and / or other optical satellite image whose water turbidity adjacent to the optical satellite image is less than or equal to the preset water turbidity threshold, determine the training sample points of the optical satellite image in the optical satellite image.

[0117] Step 603, if the geographic feature information of the target monitoring area indicates that the proportion of the flooded area is less than the preset proportion threshold, the selection range of the training sample points of the optical satellite image is determined according to the cloud shadow shape and the building shadows of the synthetic aperture radar satellite image and / or other optical satellite image at different times on adjacent dates, and the training sample points of the optical satellite image are determined within the selected range.

[0118] In an embodiment of the present application, the spatial position information includes spatial relative position information, which is used to characterize the relative position of a point in the image relative to other objects in the image. If the geographic feature information of the target monitoring area indicates that the proportion of the flooded area is less than the preset proportion threshold, the terminal uses the image other than the cloud shadow and the building shadow in the optical satellite image as the selection range of the training sample points of the optical satellite image based on the cloud shadow shape and the building shadows of the synthetic aperture radar satellite image and / or other optical satellite image at different times on adjacent dates. Then, the terminal selects the training sample points of the optical satellite image within the selection range.

[0119] In the above-mentioned monitoring method of flooding range, the extraction of small water bodies and ridges is mainly affected by the spatial resolution of the sensor. This method includes the center points of small water bodies and ridges in the training samples of lower spatial resolution images by comparing the higher spatial resolution SAR images and optical images before and after. Even if there is inevitably pixel confusion in the lower spatial resolution image, it can still basically ensure that the typical small water bodies and ridges have corresponding pixel reflections, and the flooded area is consistent with the higher spatial resolution image. The distinction between turbid water bodies and land is affected by both the sensor principle and spatial resolution. During the process of water storage and drainage, the soil and sediment will be rolled up by the faster water flow, making the water turbid. The imaging results of different sensors in this scenario are different. For some images that cannot distinguish turbid water bodies from land by visual observation, this method uses the adjacent clear images before and after as references to determine the training sample points. Due to the imaging principle of optical images, clouds and building shadows are easily misclassified as water bodies. For cloud shadows, the present invention removes them by comparing the cloud and shadow shapes. For parts of densely built areas that may be affected by building shadow images, it is mainly determined whether they are water bodies based on different times of adjacent dates (corresponding to different solar azimuths and altitudes) and higher-resolution images. The flood range sequences of different optical images in this method have good consistency and are comparable to the flooded areas extracted by SAR, which can further improve the consistency and accuracy of flood range monitoring.

[0120] In one embodiment, for optical satellite images, this method uses sensor DN (Digital Number) value data. To obtain the surface reflectance result, the terminal performs radiometric calibration, atmospheric correction, terrain correction and geometric registration. The terminal obtains radiometric calibration parameters from the sensor provided by the Natural Satellite Application Center. Then, the terminal performs radiometric calibration on the optical satellite image based on the radiometric calibration parameters. Then, the terminal uses an atmospheric correction algorithm to perform atmospheric correction on the radiometrically calibrated optical satellite image. Among them, the atmospheric correction algorithm is the FLAASH (Fast line-of-sightatmospheric analysis of spectral hypercubes) algorithm in the ENVI software. Then, the terminal applies a band fusion algorithm to the Gaofen-1 PMS image, and fuses the panchromatic band and the multispectral band to finally obtain a 2-meter resolution multispectral d surface reflectance image. Then, the terminal performs terrain correction on all optical satellite images and reprojects them to a unified projection coordinate system (UTM 50N).

[0121] In one embodiment, the target monitoring area is a total of five flood storage areas, namely B, C, D, E and F. In order to make full use of the imaging results of each satellite transit, the satellite data source for inversion of the flood range on that day is determined according to the satellite operation scheduling plan and the actually available images. Due to the high requirements for timeliness, some of the images obtained have not been preprocessed. For optical images, radiation calibration, atmospheric correction, terrain correction and geometric registration are performed in sequence. After the preprocessing is completed, the SAR satellite images and optical satellite images are classified separately. For SAR satellite images, the first image of each flood storage area during the flood period is used to calibrate all eight parameters of the generalized Gaussian mixture model of the backscattering coefficient of HH polarization (Gaofen-3 satellite) or VV polarization (Sentinel-1 satellite). For subsequent images, only the mean is adjusted to ensure the consistency of the flood range extraction results; for optical satellite images, training sample points are carefully selected and supplemented with shadow screening to obtain an optical satellite image inversion area sequence comparable to the SAR flood area sequence. For the F flood storage and detention area, because it involves the flooding of a large area of ​​farmland, special attention is paid to the extraction of field ridges when selecting samples to ensure that the field ridges that are not flooded can be identified in the optical images with lower spatial resolution. In addition, due to the large flooded area and fast flow rate of the F flood storage and detention area, attention is paid to water turbidity when selecting sample points. For the D flood storage and detention area, the E flood storage and detention area and the B flood storage and detention area, because the flooded area accounts for a small proportion of the total area of ​​the overall flood storage and detention area, the overestimation of the flooded range caused by cloud shadows will be more significant, and they are removed by comparison. At the same time, the main flooded areas of the C flood storage and detention area and the D flood storage and detention area are adjacent to the building cluster area, and the influence of building shadows is taken into account when selecting sample points. The monitoring results of the flooded area of ​​the B flood storage and detention area from July 24 to August 23 are as follows Figure 7 As shown, it includes the first real-time flooded area corresponding to the Gaofen-3 satellite image and the Sentinel-1 satellite image, and the second real-time flooded area corresponding to the Gaofen-1 satellite image, the Gaofen-6 satellite image and the Environmental Disaster Reduction-2 satellite image.

[0122] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0123] Based on the same inventive concept, the embodiment of the present application also provides a submerged range monitoring device for implementing the above-mentioned submerged range monitoring method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in the embodiments of one or more submerged range monitoring devices provided below can refer to the limitations of the submerged range monitoring method above, and will not be repeated here.

[0124] In an exemplary embodiment, Figure 8 As shown, a monitoring device 800 for flooding range is provided, comprising: a modeling module 810, a solving module 820, a first determining module 830, a calculating module 840 and a second determining module 850, wherein:

[0125] The modeling module 810 is used to establish a probability model of the backscattering coefficient of each synthetic aperture radar satellite image of the target monitoring area within the monitoring period based on the synthetic aperture radar satellite image and a preset generalized Gaussian mixture model;

[0126] A solution module 820 is used to solve the classification threshold corresponding to the synthetic aperture radar satellite image based on the probability model and the expectation maximization EM algorithm;

[0127] The first determination module 830 is used to determine the target category corresponding to each pixel based on the comparison result of the backscatter coefficient observed by each pixel in the synthetic aperture radar satellite image and the classification threshold; the target category is one of water body and non-water body;

[0128] A calculation module 840 is used to calculate a first real-time flooded area corresponding to the synthetic aperture radar satellite image according to the target category corresponding to each pixel;

[0129] The second determination module 850 is used to determine the flooded area monitoring result of the target monitoring area according to the first real-time flooded area corresponding to each of the synthetic aperture radar satellite images.

[0130] Optionally, the solution module 820 is specifically used for:

[0131] Using the expectation maximization (EM) algorithm, the probability model is calibrated to obtain calibrated parameters corresponding to the synthetic aperture radar satellite image;

[0132] Based on the calibrated parameters and a preset numerical algorithm, a classification threshold corresponding to the synthetic aperture radar satellite image is solved.

[0133] Optionally, the solution module 820 is specifically used for:

[0134] For the first synthetic aperture radar satellite image, multiple initial values ​​of the parameters to be calibrated of the probability model are tested, and a first optimal initial value set is selected; the optimal initial value set includes a first optimal initial value of a scaling parameter, a first optimal initial value of a shape parameter, a first optimal initial value of a mean, and a first optimal initial value of a category pixel ratio;

[0135] Using the expectation maximization (EM) algorithm and the optimal initial value set, the probability model is calibrated to obtain calibrated parameters corresponding to the first synthetic aperture radar satellite image;

[0136] For subsequent synthetic aperture radar satellite images within the monitoring period, the first optimal initial value of the scaling parameter, the first optimal initial value of the shape parameter, and the first optimal initial value of the category pixel ratio are used as the second optimal initial value of the scaling parameter, the second optimal initial value of the shape parameter, and the second optimal initial value of the category pixel ratio, respectively, and multiple initial values ​​of the mean of the probability model are tested to select the second optimal initial value of the mean;

[0137] The probability model is calibrated using an expectation-maximization EM algorithm, the second optimal initial value of the scaling parameter, the second optimal initial value of the shape parameter, the second optimal initial value of the mean, and the second optimal initial value of the category pixel ratio to obtain calibrated parameters corresponding to the subsequent synthetic aperture radar satellite image.

[0138] Optionally, the second determining module 850 is specifically configured to:

[0139] Acquire an optical satellite image of the target monitoring area within the monitoring period; a first image time corresponding to the synthetic aperture radar satellite image is different from a second image time corresponding to the optical satellite image;

[0140] For each optical satellite image, determining a second real-time flooded area corresponding to the optical satellite image according to the optical satellite image and a preset random forest algorithm;

[0141] The first real-time flooded area corresponding to each synthetic aperture radar satellite image and the second real-time flooded area corresponding to each optical satellite image constitute the flooded area monitoring result of the target monitoring area.

[0142] Optionally, the second determining module 850 is specifically configured to:

[0143] Determining training sample points of the optical satellite image in the optical satellite image according to the geographic feature information of the target monitoring area;

[0144] Based on the training sample points and a preset random forest algorithm, a random forest model corresponding to the optical satellite image is established;

[0145] A second real-time flooded area corresponding to the optical satellite image is determined according to the random forest model and the optical satellite image.

[0146] Optionally, the second determining module 850 is specifically configured to:

[0147] If the geographic feature information of the target monitoring area indicates the existence of small water bodies and / or field ridges, and the spatial resolution of the optical satellite image is lower than a preset resolution threshold, then according to the spatial position information of the center points of the small water bodies and / or field ridges in the synthetic aperture radar satellite images and / or other optical satellite images adjacent to the optical satellite image and having a spatial resolution higher than the preset resolution threshold, determine the center points of the small water bodies and / or field ridges in the optical satellite image, and generate training sample points including the center points of the small water bodies and / or field ridges in the optical satellite image;

[0148] If the geographic feature information of the target monitoring area indicates that there is a water flow velocity greater than a preset flow velocity threshold, and the water turbidity of the optical satellite image is greater than a preset water turbidity threshold, then according to the spatial position information of each point in the synthetic aperture radar satellite image and / or other optical satellite image whose water turbidity adjacent to the optical satellite image is less than or equal to the preset water turbidity threshold, determine the training sample points of the optical satellite image in the optical satellite image;

[0149] If the geographic feature information of the target monitoring area indicates that the proportion of the flooded area is less than a preset proportion threshold, the selection range of the training sample points of the optical satellite image is determined according to the cloud shadow shape and the building shadows of the synthetic aperture radar satellite images and / or other optical satellite images at different times on adjacent dates, and the training sample points of the optical satellite image are determined within the selection range.

[0150] Each module in the above-mentioned flood range monitoring device can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0151] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Fig. 9As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be realized through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for monitoring a flooding range is realized. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0152] Those skilled in the art will understand that Fig. 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0153] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0154] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0155] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0156] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0157] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0158] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0159] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the present application. It should be noted that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for monitoring flooding range, characterized in that: The method comprises: For each synthetic aperture radar satellite image of the target monitoring area within the monitoring period, a probability model of the backscattering coefficient of the synthetic aperture radar satellite image is established based on the synthetic aperture radar satellite image and a preset generalized Gaussian mixture model; Based on the probability model and the expectation-maximization EM algorithm, solving the classification threshold corresponding to the synthetic aperture radar satellite image; Based on the comparison result of the backscatter coefficient observed for each pixel in the synthetic aperture radar satellite image and the classification threshold, determining the target category corresponding to each pixel; the target category is one of a water body and a non-water body; Calculating a first real-time flooded area corresponding to the synthetic aperture radar satellite image according to the target category corresponding to each pixel; Determining a monitoring result of the flooded area of ​​the target monitoring area according to the first real-time flooded area corresponding to each synthetic aperture radar satellite image; Wherein, solving the classification threshold corresponding to the synthetic aperture radar satellite image based on the probability model and the expectation maximization EM algorithm includes: using the expectation maximization EM algorithm to calibrate the parameters of the probability model to obtain the calibrated parameters corresponding to the synthetic aperture radar satellite image; solving the classification threshold corresponding to the synthetic aperture radar satellite image based on the calibrated parameters and a preset numerical algorithm; the numerical algorithm is the Newton-Raphson algorithm; Among them, the use of the expectation-maximization EM algorithm to calibrate the parameters of the probability model to obtain the calibrated parameters corresponding to the synthetic aperture radar satellite image includes: for a first synthetic aperture radar satellite image, testing multiple initial values ​​of the parameters to be calibrated of the probability model, and selecting a first optimal initial value set; the first optimal initial value set includes a first optimal initial value of a scaling parameter, a first optimal initial value of a shape parameter, a first optimal initial value of a mean, and a first optimal initial value of a category pixel ratio; using the expectation-maximization EM algorithm and the first optimal initial value set to calibrate the parameters of the probability model to obtain the calibrated parameters corresponding to the first synthetic aperture radar satellite image; for subsequent synthetic aperture radars within the monitoring period, radar satellite image, taking the first optimal initial value of the scaling parameter, the first optimal initial value of the shape parameter and the first optimal initial value of the category pixel ratio as the second optimal initial value of the scaling parameter, the second optimal initial value of the shape parameter and the second optimal initial value of the category pixel ratio, respectively, and testing multiple initial values ​​of the mean of the probability model to select the second optimal initial value of the mean; using the expectation maximization EM algorithm, the second optimal initial value of the scaling parameter, the second optimal initial value of the shape parameter, the second optimal initial value of the mean and the second optimal initial value of the category pixel ratio to perform parameter calibration on the probability model to obtain the calibrated parameters corresponding to the subsequent synthetic aperture radar satellite image; Among them, determining the flooded area monitoring result of the target monitoring area according to the first real-time flooded area corresponding to each of the synthetic aperture radar satellite images includes: acquiring the optical satellite image of the target monitoring area within the monitoring period; the first image time corresponding to the synthetic aperture radar satellite image is different from the second image time corresponding to the optical satellite image; for each optical satellite image, determining the second real-time flooded area corresponding to the optical satellite image according to the optical satellite image and a preset random forest algorithm; the first real-time flooded area corresponding to each of the synthetic aperture radar satellite images and the second real-time flooded area corresponding to each of the optical satellite images constitute the flooded area monitoring result of the target monitoring area.

2. The method according to claim 1, characterized in that: The step of determining a second real-time flooded area corresponding to the optical satellite image according to the optical satellite image and a preset random forest algorithm includes: Determining training sample points of the optical satellite image in the optical satellite image according to the geographic feature information of the target monitoring area; Based on the training sample points and a preset random forest algorithm, a random forest model corresponding to the optical satellite image is established; A second real-time flooded area corresponding to the optical satellite image is determined according to the random forest model and the optical satellite image.

3. The method according to claim 2, characterized in that The step of determining, in the optical satellite image, training sample points of the optical satellite image based on the geographic feature information of the target monitoring area comprises: If the geographic feature information of the target monitoring area indicates the existence of small water bodies and / or field ridges, and the spatial resolution of the optical satellite image is lower than a preset resolution threshold, then according to the spatial position information of the center points of the small water bodies and / or field ridges in the synthetic aperture radar satellite images and / or other optical satellite images adjacent to the optical satellite image and having a spatial resolution higher than the preset resolution threshold, determine the center points of the small water bodies and / or field ridges in the optical satellite image, and generate training sample points including the center points of the small water bodies and / or field ridges in the optical satellite image; If the geographic feature information of the target monitoring area indicates that there is a water flow velocity greater than a preset flow velocity threshold, and the water turbidity of the optical satellite image is greater than a preset water turbidity threshold, then according to the spatial position information of each point in the synthetic aperture radar satellite image and / or other optical satellite image whose water turbidity adjacent to the optical satellite image is less than or equal to the preset water turbidity threshold, determine the training sample points of the optical satellite image in the optical satellite image; If the geographic feature information of the target monitoring area indicates that the proportion of the flooded area is less than a preset proportion threshold, the selection range of the training sample points of the optical satellite image is determined according to the cloud shadow shape and the building shadows of the synthetic aperture radar satellite images and / or other optical satellite images at different times on adjacent dates, and the training sample points of the optical satellite image are determined within the selection range.

4. A flood range monitoring device, characterized in that: The device comprises: A modeling module is used to establish a probability model of the backscattering coefficient of each synthetic aperture radar satellite image of the target monitoring area within the monitoring period based on the synthetic aperture radar satellite image and a preset generalized Gaussian mixture model; A solution module, used for solving the classification threshold corresponding to the synthetic aperture radar satellite image based on the probability model and the expectation maximization EM algorithm; A first determination module is used to determine the target category corresponding to each pixel based on the comparison result of the backscatter coefficient observed by each pixel in the synthetic aperture radar satellite image and the classification threshold; the target category is one of water body and non-water body; A calculation module, used for calculating a first real-time flooded area corresponding to the synthetic aperture radar satellite image according to the target category corresponding to each pixel; A second determination module is used to determine the flooded area monitoring result of the target monitoring area according to the first real-time flooded area corresponding to each synthetic aperture radar satellite image; The solution module is specifically used to: use the expectation maximization EM algorithm to calibrate the parameters of the probability model to obtain the calibrated parameters corresponding to the synthetic aperture radar satellite image; based on the calibrated parameters and a preset numerical algorithm, solve the classification threshold corresponding to the synthetic aperture radar satellite image; the numerical algorithm is the Newton-Raphson algorithm; The solution module is specifically used to: for the first synthetic aperture radar satellite image, test multiple initial values ​​of the parameters to be calibrated of the probability model, and select a first optimal initial value set; the first optimal initial value set includes the first optimal initial value of the scaling parameter, the first optimal initial value of the shape parameter, the first optimal initial value of the mean, and the first optimal initial value of the category pixel ratio; use the expectation maximization EM algorithm and the first optimal initial value set to calibrate the parameters of the probability model to obtain the calibrated parameters corresponding to the first synthetic aperture radar satellite image; for subsequent synthetic aperture radar satellite images within the monitoring period, the first optimal initial value of the scaling parameter is used as the initial value of the first optimal initial value set. The first optimal initial value of the shape parameter and the first optimal initial value of the proportion of the category pixels are respectively used as the second optimal initial value of the scaling parameter, the second optimal initial value of the shape parameter and the second optimal initial value of the proportion of the category pixels, and multiple initial values ​​of the mean of the probability model are tested to select the second optimal initial value of the mean; the expectation maximization EM algorithm, the second optimal initial value of the scaling parameter, the second optimal initial value of the shape parameter, the second optimal initial value of the mean and the second optimal initial value of the proportion of the category pixels are used to calibrate the parameters of the probability model to obtain the calibrated parameters corresponding to the subsequent synthetic aperture radar satellite image; Among them, the second determination module is specifically used to: obtain the optical satellite image of the target monitoring area within the monitoring period; the first image time corresponding to the synthetic aperture radar satellite image is different from the second image time corresponding to the optical satellite image; for each optical satellite image, determine the second real-time flooded area corresponding to the optical satellite image according to the optical satellite image and a preset random forest algorithm; the first real-time flooded area corresponding to each synthetic aperture radar satellite image and the second real-time flooded area corresponding to each optical satellite image constitute the flooded area monitoring result of the target monitoring area.

5. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

7. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

Images