A method and system for basin flood risk assessment based on particle recognition algorithm

Abstract

The present invention discloses a basin flood risk assessment method and system based on a particle recognition algorithm, and belongs to the field of flood warning technology. Although the existing scheme can measure the real-time flow of the river, it does not involve how to deduce and evaluate the flood inundation situation, resulting in the inability to quickly know the flood inundation situation around the river bank. The basin flood risk assessment method based on a particle recognition algorithm of the present invention, by constructing a basin spatial discrete model, a flow velocity calculation model, a flow simulation model, a two-dimensional hydrodynamic model, and a flood risk assessment model, can not only calculate the cross-sectional flow of the basin, but also deduce and evaluate the flood inundation situation, so that the flood inundation risk and inundation data around the water area can be quickly known, so as to meet the needs of basin flood risk analysis, improve the accuracy of flood disaster warning, and provide more data supplement and management basis for basin flood control.

Description

Technical Field

[0001] The invention relates to a basin flood risk assessment method and system based on a particle recognition algorithm, belonging to the technical field of flood early warning. Background Art

[0002] The Chinese patent application (publication number: CN117935486A) discloses a flood warning method and system for wild rivers in arid areas, including: selecting a suitable position on the riverbank of the wild river to arrange a camera, and placing a water gauge near a fixed reference object in the wild river; using the camera to shoot the surface of the wild river in real time and collect the real-time rainfall in the arid area; extracting a single frame image from the real-time video, preprocessing each frame of the image, and obtaining a wild river image; recognizing the image, and using the water gauge information in the image to obtain the water depth and river width of the wild river, using the particle image velocimetry method to calculate the real-time flow velocity on the surface of the wild river, and calculating the real-time flow of the wild river according to the water depth, river width and real-time flow velocity; issuing a warning twice according to the relationship between the real-time rainfall, the real-time flow and the warning index.

[0003] Although the above scheme can measure the real-time flow of the river, it does not involve how to deduce and evaluate the flood inundation situation, resulting in the inability to quickly know the flood inundation risk and inundation data around the river bank, and therefore cannot meet the needs of basin flood risk analysis.

[0004] The information disclosed in this Background Art is only for understanding the background of the inventive concept and therefore it may include information that does not constitute the prior art. Summary of the invention

[0005] In response to the above problem or one of the above problems, an object of the present invention is to provide a basin flood risk assessment method and system based on a particle recognition algorithm. By constructing a basin spatial discrete model, a flow velocity calculation model, a flow simulation model, a two-dimensional hydrodynamic model, and a flood risk assessment model, not only can the cross-sectional flow of the basin be calculated, but also the flood inundation situation can be deduced and evaluated, so that the flood inundation risk and inundation data around the water area can be quickly known, thereby meeting the needs of basin flood risk analysis and improving the accuracy of flood disaster warning.

[0006] In response to the above problem or one of the above problems, the second object of the present invention is to provide a basin flood risk assessment method and system based on a particle recognition algorithm, which combines the flow measurement results with a two-dimensional hydrodynamic simulation model to deduce the flood inundation situation under the current flow and water level conditions, thereby achieving an overall grasp of the flood risk distribution of the entire basin and the flood hazard of the entire basin, and providing more data supplements and management basis for basin flood control.

[0007] To achieve one of the above purposes, the first technical solution of the present invention is:

[0008] A basin flood risk assessment method based on particle recognition algorithm includes the following contents:

[0009] Through the pre-built watershed spatial discrete model, a watershed and surrounding terrain are modeled to obtain a watershed simulation object;

[0010] Using the pre-built velocity calculation model, one or more groups of water flow images of the watershed simulation object are obtained, and based on the particle image recognition algorithm, one or more groups of water flow images are serially analyzed to calculate the velocity field information of the watershed;

[0011] Using the pre-built flow simulation model, based on the basin section location and section elevation data of the basin simulation object, the flow velocity field information of the basin is calculated to obtain one or more section flows;

[0012] According to the pre-built two-dimensional hydrodynamic model, based on the cross-sectional flow and water level data of the basin simulation object, the flood inundation situation of the basin simulation object is simulated to obtain the flood change simulation result;

[0013] Through the pre-built flood risk assessment model, the flood change simulation results are evaluated, the flood risk distribution of the basin simulation object is obtained, and the basin flood risk assessment based on the particle recognition algorithm is completed.

[0014] The present invention constructs a watershed spatial discrete model, a flow rate calculation model, a flow simulation model, a two-dimensional hydrodynamic model, and a flood risk assessment model. It can not only calculate the cross-sectional flow of the watershed, but also deduce and evaluate the flood inundation situation, so as to quickly know the flood inundation risk and inundation data around the water area. Therefore, it can meet the needs of watershed flood risk analysis and improve the accuracy of flood disaster warning.

[0015] Furthermore, the present invention combines the flow measurement results with the two-dimensional hydrodynamic simulation model, which can deduce the flood inundation situation under the current flow and water level conditions, thereby achieving an overall grasp of the flood risk distribution and flood hazard of the entire river basin, and providing more data supplements and management basis for river basin flood control.

[0016] Furthermore, the present invention performs simulation modeling based on a particle image recognition algorithm and can directly use naturally existing splashes, foam or fallen leaves on the surface of the water flow as tracer particles for flow measurement. It can be applied to water scenes with different flow velocities and flow states, has a wider scope of application, and the solution is scientific, reasonable and feasible.

[0017] As the preferred technical measures:

[0018] The method of modeling a watershed and its surrounding terrain through the pre-constructed watershed spatial discrete model to obtain the watershed simulation object is as follows:

[0019] Get the scope of a watershed to be evaluated;

[0020] According to the scope to be assessed, obtain elevation data, basin section location and water level information of several locations;

[0021] Based on the elevation data and the location of the watershed section, a two-dimensional grid divided in an unstructured manner is constructed;

[0022] Based on the terrain characteristics, the two-dimensional grid is locally refined to represent the topographical changes of the embankment and / or villages and towns;

[0023] Perform linear interpolation on the water level information to form water level data, and assign it to the two-dimensional grid as the water level condition;

[0024] According to the conditions of the underlying surface of the watershed and different areas on the shore, several roughness coefficients are set;

[0025] The water level condition, roughness coefficient and two-dimensional grid are coupled to obtain the watershed simulation object.

[0026] As the preferred technical measures:

[0027] Using the pre-built velocity calculation model, one or more groups of water flow images of the watershed simulation object are obtained, and based on the particle image recognition algorithm, one or more groups of water flow images are serially analyzed to calculate the watershed velocity field information as follows:

[0028] Constructing a water flow image acquisition unit for acquiring one or more groups of water flow images of a certain watershed;

[0029] Constructing a velocity analysis unit for serially analyzing one or more groups of water flow images and calculating the instantaneous velocity field data of water flow at a number of simulation grid points;

[0030] Decompose the instantaneous velocity field data of water flow into lateral velocity and longitudinal velocity;

[0031] According to the value range limited by the flow velocity direction, the lateral velocity and the longitudinal velocity are processed, and the instantaneous velocity field data of the water flow that does not conform to the flow velocity direction are eliminated to obtain the first instantaneous velocity field data of the water flow;

[0032] According to the minimum coefficient threshold of the cross-correlation, the first water flow instantaneous velocity field data is filtered to obtain the second water flow instantaneous velocity field data;

[0033] According to the position of the simulation grid points, the instantaneous velocity field data of the second water flow are averaged to calculate the two-dimensional average velocity field of the basin surface corresponding to the video shooting period, that is, the basin flow velocity field information.

[0034] As the preferred technical measures:

[0035] The method of constructing the water flow image acquisition unit is as follows:

[0036] Based on the statistical frequency of hydrological data in the basin and the demand for points, determine the required flow measurement points in the basin and count the number of flow measurement points;

[0037] Set the video collection duration information based on the frequency of hydrological data collection;

[0038] According to the video acquisition duration information, the video acquisition data of the flow measurement point is obtained, and the resolution and frame rate of the camera are obtained;

[0039] Perform image acquisition on the video acquisition data, retain one image for every n images, and number the images in chronological order to form an initial water flow image group;

[0040] Based on the distance between the markers on both sides of the watershed in the water flow image, the images in the initial water flow image group are orthorectified to restore the image's scale and spatial geographic coordinate information in reality, and obtain one or more groups of orthorectified orthographic image groups;

[0041] The top view image group is used as the water flow image of a certain river basin to complete the construction of the water flow image acquisition unit;

[0042] The method of constructing a speed analysis unit is as follows:

[0043] According to the required calculation accuracy of the two-dimensional velocity field and the width of a certain watershed, the required number of simulation grid points is calculated;

[0044] The number of simulation grid points includes the number of horizontal grid points and the number of vertical grid points;

[0045] According to the number of horizontal grid points and the number of vertical grid points, the orthorectified water flow image is grid-divided to obtain a number of simulation grid points;

[0046] Based on the simulation grid points on two continuous water flow images, a correlation calculation unit is constructed to obtain the surface two-dimensional instantaneous velocity field between the two continuous water flow images;

[0047] Utilizing the correlation calculation unit, serial correlation calculation is performed on all water flow images to obtain the two-dimensional surface instantaneous velocity field between the water flow images, namely, the instantaneous velocity field data of the water flow, thus completing the construction of the velocity analysis unit.

[0048] As the preferred technical measures:

[0049] Based on the distance between the landmarks on both sides of the watershed in the water flow image, the method for orthorectifying the images in the initial water flow image group is as follows:

[0050] When the image is taken at a top-down angle, the scale calibration method is used for orthorectification, which includes the following:

[0051] Obtain two mutually perpendicular landmarks in the initial water flow image group;

[0052] For two mutually perpendicular markers, connect them to obtain the physical distance and pixel distance between the two markers;

[0053] Determine the horizontal and vertical proportional relationship between physical distance and pixel distance;

[0054] All images are scaled in proportion to obtain one or more groups of orthographic image groups after orthorectification;

[0055] When the image is taken at an oblique viewing angle, the four-point calibration method is used for orthorectification, which includes the following:

[0056] Select four markers in the initial water flow image group, any three of which are not collinear;

[0057] Calculate the physical distances between four landmarks;

[0058] According to the physical distance, the relative physical coordinates of the four markers are determined and used as the coordinates of the calibration points;

[0059] Based on the calibration point coordinates, the conversion coefficient between the image physical coordinates and the pixel coordinates is obtained;

[0060] Apply the conversion coefficient to each pixel in the image to complete the orthorectification of the image;

[0061] The orthorectified image is cropped in a rectangular shape to retain the area where the flow velocity and flow rate need to be calculated in the water flow image, so as to obtain one or more groups of orthographic image groups;

[0062] Or / and, based on the simulation grid points on two continuous water flow images, a correlation calculation unit is constructed to obtain the surface two-dimensional instantaneous velocity field between the two continuous water flow images as follows:

[0063] Create a comparison area 1 with a simulation grid point in a water flow image as the center;

[0064] According to the serial number of a water flow image, adjacent water flow images are obtained;

[0065] Create a second comparison area centered on the simulated grid points of the adjacent water flow images;

[0066] Obtain pixel values ​​and pixel intensity distributions of contrast area 1 and contrast area 2 respectively, and set a search range;

[0067] Based on the search range, pixel value and pixel intensity distribution, the correlation between the comparison area 1 and the comparison area 2 is calculated to obtain a correlation coefficient;

[0068] According to the correlation coefficient, the displacement of the simulation grid points within a certain time period is calculated;

[0069] According to the displacement and the corresponding time period, the two-dimensional instantaneous velocity field of the surface between two consecutive water flow images is obtained.

[0070] As the preferred technical measures:

[0071] Using the pre-built flow simulation model, based on the basin section location and section elevation data of the basin simulation object, the flow velocity field information of the basin is calculated to obtain the cross-sectional flow as follows:

[0072] Based on the location and elevation data of the watershed section and the velocity field information of the watershed, the surface velocity on a watershed section is obtained;

[0073] The conversion coefficient is calculated based on the roughness of the basin where the watershed section is located, the water velocity at a certain depth, and the ratio of a certain depth to the water depth of the section;

[0074] According to the conversion coefficient, the surface velocity of the watershed section is converted into the average velocity of the cross-section perpendicular to the cross-section;

[0075] The cross-sectional flow rate is calculated by the average flow velocity of the cross-sectional vertical line and the cross-sectional water depth, combined with the width and depth of each cross-sectional vertical line.

[0076] As the preferred technical measures:

[0077] Based on the basin section location and section elevation data, and according to the basin velocity field information, the method for obtaining the surface velocity on the basin section is as follows:

[0078] Based on the location of the watershed section, a simulation straight line where the section is located is constructed;

[0079] On the simulation line, multiple calculation points are evenly set;

[0080] According to the cross-section elevation data, the interval distance value between the calculation points is obtained;

[0081] Take the calculation point as the center of the watershed circle, set the interval distance value as the search radius, and get the search area;

[0082] Searching for existing simulation grid points in the search area, obtaining a number of simulation grid points and corresponding flow velocity field information of the watershed;

[0083] Averaging the velocity field information of several watersheds to obtain the average velocity field information;

[0084] The average velocity field information is used as the surface velocity of the calculation point and assigned to the corresponding calculation point, that is, the center of the basin;

[0085] Based on the center of the basin and the corresponding average flow velocity field information, linear interpolation is performed on the calculation points on the simulation line to obtain the surface velocity of each calculation point on the basin section.

[0086] As the preferred technical measures:

[0087] According to the pre-built two-dimensional hydrodynamic model, based on the cross-sectional flow and water level data of the basin simulation object, the flood inundation situation of the basin simulation object is simulated, and the method for obtaining the flood change simulation result is as follows:

[0088] Obtain the hydrodynamic data of the basin simulation object, including water depth, water surface elevation, sampling period, average water velocity value, basin bottom friction coefficient, surface wind stress, Coriolis force, atmospheric pressure and turbulent viscosity diffusion coefficient;

[0089] Based on the hydrodynamic data, the continuity equation and momentum equation are established to form a two-dimensional hydrodynamic equation group;

[0090] The cross-sectional flow and water level data of the basin simulation object are input into the two-dimensional hydrodynamic equations as initial conditions and boundaries, and the basin simulation calculation is performed to obtain the two-dimensional water depth field, water level field and flow velocity field of the basin;

[0091] Based on the two-dimensional water depth field and water level field of the basin, combined with the surface elevation data, the water depth data belonging to the water area is eliminated to obtain the flooded water depth data;

[0092] According to the flooded water depth data, grid nodes with water depth greater than or equal to 0.001 meters are counted as flooded nodes, and the flooded nodes with water depth equal to 0.001 meters are connected to obtain contour lines;

[0093] The range surrounded by the contour lines is regarded as the flooded area; and the area of ​​the flooded area is calculated as the flooded area;

[0094] Based on the velocity field, the velocity field in the water area and the velocity field on land are distinguished according to the water shoreline, and the velocity in the flooded area and the watershed section exceeding the velocity safety range are obtained;

[0095] Based on the inundated area, inundated surface, inundated water depth data, flow velocity in the inundated area and the basin section exceeding the flow velocity safety range, the flood inundation situation of the basin simulation object is determined and the flood change simulation results are obtained.

[0096] As the preferred technical measures:

[0097] The method of conducting risk assessment on the flood change simulation results through the pre-built flood risk assessment model and obtaining the flood risk distribution of the basin simulation object is as follows:

[0098] According to the flood range and flood area, a flood inundation map of the basin simulation object is produced to obtain the areas with flooding risks under the current flood state;

[0099] According to the area where there is a risk of flooding, obtain the flooding depth and flow rate of the area;

[0100] Multiply the flood depth and flow velocity to obtain the flood risk assessment index;

[0101] According to the inundation risk assessment index and the basin flood hazard level classification standard, the inundated area is divided into hazard levels, and the current flood hazard level map is drawn to obtain the flood risk distribution of the basin simulation object under the current flood state.

[0102] To achieve one of the above purposes, the second technical solution of the present invention is:

[0103] A basin flood risk assessment system based on particle recognition algorithm, comprising:

[0104] one or more processors;

[0105] A storage device for storing one or more programs;

[0106] When the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned basin flood risk assessment method based on particle recognition algorithm.

[0107] Compared with the prior art solutions, the present invention has the following beneficial effects:

[0108] The present invention constructs a watershed spatial discrete model, a flow rate calculation model, a flow simulation model, a two-dimensional hydrodynamic model, and a flood risk assessment model. It can not only calculate the cross-sectional flow of the watershed, but also deduce and evaluate the flood inundation situation, so as to quickly know the flood inundation risk and inundation data around the water area. Therefore, it can meet the needs of watershed flood risk analysis and improve the accuracy of flood disaster warning.

[0109] Furthermore, the present invention combines the flow measurement results with the two-dimensional hydrodynamic simulation model, which can deduce the flood inundation situation under the current flow and water level conditions, thereby achieving an overall grasp of the flood risk distribution and flood hazard of the entire river basin, and providing more data supplements and management basis for river basin flood control.

[0110] Furthermore, the present invention performs simulation modeling based on a particle image recognition algorithm and can directly use naturally existing splashes, foam or fallen leaves on the surface of the water flow as tracer particles for flow measurement. It can be applied to water scenes with different flow velocities and flow states, has a wider scope of application, and the solution is scientific, reasonable and feasible. BRIEF DESCRIPTION OF THE DRAWINGS

[0111] Figure 1 A schematic diagram of a flow chart of a basin flood risk assessment method of the present invention;

[0112] Figure 2 A schematic diagram of calculating the two-dimensional average velocity according to the present invention. DETAILED DESCRIPTION

[0113] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention 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 invention and are not intended to limit the present invention.

[0114] On the contrary, the present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention as defined by the claims. Further, in order to make the public have a better understanding of the present invention, some specific details are described in detail in the following detailed description of the present invention. Those skilled in the art can fully understand the present invention without the description of these details.

[0115] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention pertains. The terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "or / and" as used herein includes any and all combinations of one or more of the associated listed items.

[0116] like Figure 1 As shown, the first specific embodiment of the basin flood risk assessment method based on the particle recognition algorithm of the present invention is:

[0117] A basin flood risk assessment method based on particle recognition algorithm includes the following contents:

[0118] Through the pre-built watershed spatial discrete model, a watershed and surrounding terrain are modeled to obtain a watershed simulation object;

[0119] Using the pre-built velocity calculation model, one or more groups of water flow images of the watershed simulation object are obtained, and based on the particle image recognition algorithm, one or more groups of water flow images are serially analyzed to calculate the velocity field information of the watershed;

[0120] Using the pre-built flow simulation model, based on the basin section location and section elevation data of the basin simulation object, the flow velocity field information of the basin is calculated to obtain one or more section flows;

[0121] According to the pre-built two-dimensional hydrodynamic model, based on the cross-sectional flow and water level data of the basin simulation object, the flood inundation situation of the basin simulation object is simulated to obtain the flood change simulation result;

[0122] Through the pre-built flood risk assessment model, the flood change simulation results are evaluated, the flood risk distribution of the basin simulation object is obtained, and the basin flood risk assessment based on the particle recognition algorithm is completed.

[0123] The second specific embodiment of the basin flood risk assessment method based on the particle recognition algorithm of the present invention:

[0124] A basin flood risk assessment method based on particle recognition algorithm includes the following steps:

[0125] Step 1, obtaining one or more groups of water flow images of a certain river basin through a pre-built water flow image acquisition unit;

[0126] Step 2: using a pre-built velocity analysis unit, based on a particle image recognition algorithm, serially analyzing one or more groups of water flow images, and calculating the instantaneous velocity field data of the water flow;

[0127] Step 3: Using the pre-built velocity calculation model, the instantaneous velocity field data of the water flow is processed according to the velocity direction and particle correlation to obtain the velocity field information of the watershed;

[0128] Step 4: using a pre-built flow simulation model, based on the location and elevation data of the watershed section, calculate the flow velocity field information of the watershed to obtain one or more section flows;

[0129] Step five: Through the pre-constructed flood risk assessment model, combined with the basin velocity field information, conduct a risk assessment on the flow of one or more sections to obtain the flood risk distribution of the basin and complete the basin flood risk assessment based on the particle recognition algorithm.

[0130] The watershed refers to the water area and the surrounding land and buildings, and the water area refers to river water, dam water or lake water.

[0131] The third specific embodiment of the basin flood risk assessment method based on the particle recognition algorithm of the present invention is:

[0132] A basin flood risk assessment method based on particle image recognition algorithm includes the following steps:

[0133] S1: Based on the statistical frequency and location requirements of the hydrological data in the basin, use the camera at each flow measurement point to collect video at a fixed time frequency. Determine the video resolution and frame rate through the captured video of the river flow in the basin, and collect images of the video to obtain a group of images with the same time interval and numbered in chronological order.

[0134] S2: Based on the distance between the landmarks on both sides of the river in the image, all the collected images are orthorectified to restore the image's scale and spatial geographic coordinate information in reality, and obtain a group of orthorectified river channel top-view images.

[0135] S3: Based on the particle image recognition algorithm, all images in the image group are serially analyzed to calculate the two-dimensional instantaneous velocity field data corresponding to the river surface in a group of images.

[0136] S4: Based on the obtained two-dimensional instantaneous velocity field data, the results are filtered using the thresholds of velocity magnitude and direction and particle correlation, abnormal velocities are eliminated, and then the filtered two-dimensional instantaneous velocity field is averaged to obtain the two-dimensional flow velocity field of the river in the corresponding period of the video.

[0137] S5: Based on the measured river section position and section elevation data, combined with the water level information during the video shooting period and the riverbed roughness of the river, the surface flow velocity is converted into the average flow velocity of the section vertical line, and the section flow at the section location is calculated.

[0138] S6: Based on the flood risk classification standard of the basin, the risk level of the obtained two-dimensional velocity field and cross-sectional flow results of the river is judged, and the velocity and flow results and risk level of each flow measuring point are counted.

[0139] S7: Based on the calculated cross-sectional flow and collected water level results, combined with the two-dimensional hydrodynamic model, the current flood inundation situation is simulated.

[0140] S8: Based on the results of flood inundation simulation, statistical analysis is conducted on the inundation range, area, inundation depth, flow velocity in the inundated area, and river sections that exceed the safe flow velocity range to obtain the real-time distribution of flood risks in the entire river basin, providing a basis for flood control management in the river basin.

[0141] In this embodiment, S1, based on the statistical frequency of hydrological data in the basin and the requirements of the points, the camera at each flow measurement point is used to collect video at a fixed time frequency, the video resolution and frame rate are determined by taking the water flow video of the river in the basin, and the video is collected to obtain a group of images with the same time interval and numbered in chronological order. The specific steps are as follows:

[0142] S1-1: Determine the required flow measurement points in the basin and count the number of flow measurement points, which is Through the camera that captures the river flow at each flow measurement point, the camera can be an existing river monitoring camera or a newly deployed camera. , every Time collection for a period of time For video, usually set 10-60s, and get the camera resolution and frame rate at the same time .

[0143] S1-2: Capture images from the video. To increase the calculation speed, you can select One image is retained, and a total of An image, whose expression is as follows:

[0144]

[0145]

[0146] Number the images in chronological order to get the image numbers The time interval between two images in sequence in the image group is Seconds, the time interval is calculated as follows:

[0147]

[0148] In this embodiment, S2, based on the distance between the landmarks on both sides of the river in the image, orthorectifies all the collected images, restores the image's scale and spatial geographic coordinate information in reality, and obtains the orthorectified river channel top view image group. The specific steps are as follows:

[0149] S2-1: Orthorectify the collected images. If the shooting angle is already a top-down view, the scale calibration method can be used, that is, The physical distance and pixel distance of the two pairs of perpendicular lines in the image are determined to determine the horizontal and vertical proportional relationship between the physical distance and the pixel distance. , scaled according to the proportional relationship to obtain the image after orthorectification according to the actual proportional relationship .

[0150] If the shooting angle is oblique, the four-point calibration method can be used, that is, Four markers The relative physical coordinates of the four markers can be determined by the physical distance between (any three non-collinear) points, or the physical coordinates of the four markers can be directly obtained as the coordinates of the calibration points to obtain the conversion coefficient between the image physical coordinates and the pixel coordinates, which is then applied to each pixel point of the image to complete the orthorectification of the image.

[0151] Afterwards, the image Perform the same operation to obtain the orthorectified image .

[0152] S2-2: Perform rectangular cropping on the orthorectified top-view image to retain the river area in the image where the flow velocity and flow rate need to be calculated.

[0153] In this embodiment, S3, based on the particle image recognition algorithm, serially analyze all images in the image group to calculate the two-dimensional instantaneous velocity field data corresponding to the river surface in a group of images. The specific steps are as follows:

[0154] S3-1: Based on orthorectified images , and divide the simulation grid points.

[0155] According to the required calculation accuracy of the 2D velocity field and the width of the river channel, set the image range horizontally. Simulation grid points, vertical setting simulation grid points, each simulation grid point is express.

[0156] S3-2: Based on the particle image recognition algorithm, the images in the orthorectified image group are sequentially cross-correlated. Create a contrasting area for the center , and the next figure with point The cross-correlation calculation is performed on the same contrast area created for the center, and the time interval between the two images is Second.

[0157] Calculate the correlation coefficient The expression is as follows:

[0158]

[0159] in, and is the size of the comparison area in pixels, is the pixel intensity distribution in the first query region, is the pixel intensity distribution in the second query region, yes The sample mean of yes The sample mean of .

[0160] For the simulation grid points Set a search range for the possible locations in the next image. Perform cross-correlation calculations. Iterate the calculations for each simulation grid point in the entire image. The closer the pixel intensity distribution in the contrasting area, the stronger the cross-correlation, so the particle image recognition algorithm works best when there are tracer particles with uniform distribution and relatively fixed shapes on the surface of the river water, such as foam, water spray or fallen leaves.

[0161] exist During this time, the water flows from The most likely displacement is the corresponding cross-correlation The maximum displacement. Therefore, the instantaneous 2D velocity field on the surface between two consecutive images is , by calculating the Pixel displacement Divide by time The two-dimensional instantaneous velocity field is obtained The calculation formula is as follows:

[0162]

[0163] S3-3: Serially perform the calculations in S3-2 on the entire orthorectified image set.

[0164] Through the image pair ( ),( ), ...( ) performs the calculation in S3-2 to obtain the two-dimensional surface instantaneous velocity field data of group X-1.

[0165] In this embodiment, S4, based on the obtained two-dimensional instantaneous velocity field data, the results are filtered using the thresholds of velocity magnitude and direction and the particle cross-correlation, abnormal velocities are eliminated, and then the filtered two-dimensional instantaneous velocity field is averaged to obtain the specific steps of the two-dimensional flow velocity field of the river channel in the corresponding time period of the video as follows:

[0166] S4-1: Filter the two-dimensional instantaneous velocity field data of each image. Since the shooting angle of each flow measurement point is fixed, the abnormal values ​​of the two-dimensional instantaneous velocity field can be filtered by the main direction of the flow velocity and the threshold of the cross-correlation coefficient.

[0167] The two-dimensional instantaneous velocity field of each simulation grid point Disassembled into ( ; ), get the speed in the x and y directions, and limit it according to the main direction of the flow velocity ( ; ) value range, and remove the result points that do not conform to the mainstream velocity direction; for the cross-correlation, set the minimum coefficient threshold to filter out the velocity result points whose cross-correlation is lower than the threshold, and finally obtain the two-dimensional instantaneous velocity field of each image .

[0168] S4-2: Average all the two-dimensional instantaneous velocity fields. According to the position of the simulation grid point, calculate the average velocity of each simulation grid point, so as to obtain the two-dimensional average velocity field of the river surface corresponding to the video shooting period. , which is expressed as follows:

[0169]

[0170] in, is the number of images, For the The two-dimensional instantaneous velocity field of an image.

[0171] In this embodiment, S5, based on the measured river section position and section elevation data, combined with the water level information during the video shooting period and the riverbed roughness of the river, the surface flow velocity is converted into the average flow velocity of the cross-section vertical line, and the specific steps of calculating the cross-section flow at the cross-section location are as follows:

[0172] S5-1: Calculate the surface velocity of the river section. Set multiple calculation points evenly along the straight line where the section is located. , the interval between each calculation point is , at each calculation point Search for existing simulation grid points within the radius, average the two-dimensional instantaneous velocity results of the simulation grid points, assign them to the corresponding calculation points (center of the circle), and then perform linear interpolation on the calculation point data on the straight line where the section is located to obtain the new surface velocity results for each calculation point in the river section. .

[0173] S5-2: Calculate the average vertical velocity of the cross section. Converted into cross-section vertical average velocity , and its calculation formula is as follows:

[0174]

[0175] in, is the conversion coefficient, which is mainly related to the roughness and water depth of the river where the section is located. The calculation formula is as follows:

[0176]

[0177]

[0178] in, yes The velocity of the water at depth, is the current depth, is the cross-sectional water depth, is the roughness correlation coefficient.

[0179] Through the historical measured data of the section, combined with the above formula, the roughness correlation coefficient can be obtained or through experience combined with the riverbed roughness Make a judgment to obtain the conversion coefficient , calculate the average flow velocity of the vertical line of the cross section .

[0180] S5-3: Calculate cross-sectional flow.

[0181] The average flow velocity of each vertical line of the cross section obtained by S5-2, combined with the width of each vertical line sample strip and depth , to calculate the cross-sectional flow rate , and its calculation formula is as follows:

[0182]

[0183]

[0184] in, for depth; The water depth of the cross section can be obtained in real time through cameras with water gauge recognition, radar water level meters, etc. After measuring the depth of the section, The interpolation of the calculation points, so far the cross-sectional flow of the river section in the basin during the shooting period is obtained .

[0185] In this embodiment, S6, based on the flood risk classification standard of the river basin, the risk level of the obtained two-dimensional velocity field and cross-sectional flow results of the river channel is judged, and the specific steps of calculating the velocity and flow results and risk level of each flow measurement point are as follows:

[0186] S6-1: Count the maximum flow velocity, average flow velocity and cross-sectional flow rate of all flow measurement points in this period. By calculating the average flow velocity of all cross-sections The maximum value of is obtained, and its expression is as follows:

[0187]

[0188] Average flow rate By calculating the average of the vertical average flow velocity of all sections get.

[0189] S6-2: Combined with the basin flood risk classification standards, the maximum flow velocity, average flow velocity and cross-sectional flow are divided into different risk levels. The classification results and flow measurement results of the flow measurement points in the entire basin are integrated to obtain a real-time flood risk distribution point map of the entire basin.

[0190] In this embodiment, S7, based on the calculated cross-sectional flow and the collected water level results, combined with the two-dimensional hydrodynamic model, the specific steps of deducing the current flood inundation situation are as follows:

[0191] S7-1: Based on the monitoring scope of the basin, combined with the refined elevation data and river section elevation data, a two-dimensional grid divided in an unstructured manner is constructed.

[0192] When there is a key study area in the basin, the grid in this area needs to be locally encrypted. The key study area is often the area where the key risk villages are located in the basin. The encryption usually needs to reach a level that can clearly show the terrain characteristics of the area.

[0193] S7-2: The flow and water level data obtained from the flow measurement points are used as the initial conditions and boundary inputs of the two-dimensional hydrodynamic model. The basin is simulated based on the two-dimensional hydrodynamic equations, and the two-dimensional water depth field, water level field, and flow velocity field of the basin are output.

[0194] Among multiple flow measuring points, the flow measuring point closest to the upstream boundary of the river in the basin is selected, and the obtained flow and water level results are used as the upstream boundary conditions of the corresponding river section in the model.

[0195] The water level data collected from the flow measurement point closest to the downstream boundary of the basin are selected as the downstream boundary conditions of the corresponding river section in the model. The flow and water level results obtained from the remaining flow measurement points are used for model calibration.

[0196] During the simulation calculation process, the two-dimensional hydrodynamic model can use a two-dimensional hydrodynamic equation group for solving the shallow water equation, which includes the continuity equation and the momentum equation.

[0197] The continuity equation is expressed as follows:

[0198]

[0199] The momentum equation is expressed as follows:

[0200]

[0201]

[0202] In the formula, is the water depth in meters ; is the surface elevation; is the time, the unit is second; and The lower edge of the Cartesian coordinate system , The average speed value of the direction; is the acceleration due to gravity; and are the source terms of the equations, including bottom friction, surface wind stress, Coriolis force, atmospheric pressure, etc.; is the turbulent viscous diffusion coefficient, is the gradient function, is the divergence function.

[0203] In this embodiment, S8, based on the flood inundation simulation results, statistically analyze the inundation range, area, inundation depth, flow velocity in the inundated area, and river sections that exceed the flow velocity safety range to obtain the real-time distribution of flood risks in the entire river basin, and provide a basis for flood control management in the river basin. The specific steps are as follows:

[0204] S8-1: Extract the water depth field from the simulation results of the two-dimensional hydrodynamic model and conduct statistical analysis of the inundation range, area, and inundation depth in combination with the surface elevation.

[0205] The water depth field data after the simulation results are stabilized is intercepted, and the result data belonging to the river channel is eliminated according to the river bank line, and the water depth statistics of the remaining result data are performed. The grid nodes with a water depth greater than or equal to 0.001m in the two-dimensional water depth results are counted as submerged nodes, and the submerged nodes with a water depth equal to 0.001m are connected, and contour lines are drawn as the submerged range. According to the delineated submerged range, the area of ​​the grid within the range is counted as the submerged area. The water depth results of the submerged area outside the river channel of the entire basin are divided into different levels according to the threshold, and contour lines are drawn.

[0206] S8-2: Extract the velocity field from the simulation results of the two-dimensional hydrodynamic model and conduct statistical analysis of the velocity in the flooded area and the river section that exceeds the safety range of velocity.

[0207] The velocity field data after the simulation results are stabilized are intercepted, and the result data belonging to the river range and the land results are distinguished according to the river bank line. Combined with the statistical results of S8-1, the velocity results are divided into different levels according to the threshold value for the flooded area range, and contour lines are drawn. The velocity result data belonging to the river range are graded according to the threshold value defined by the flood hazard level of the basin, and the range of river velocity exceeding the safety threshold is delineated by contour lines.

[0208] S8-3: Based on the statistical analysis results and the basin flood risk classification standards, the flooded areas are comprehensively classified by flood depth and flow velocity, and the real-time distribution of flood risks in the entire basin and the flood risk distribution point map are displayed, which specifically includes the following contents:

[0209] First, based on the inundation range and inundation area of ​​the delineated river basin, a current flood inundation map of the entire river basin is prepared to provide areas in the river basin that may be at risk of inundation under the current flood status.

[0210] According to the flood risk level classification standard of the basin, the flood depth and flow rate are used as evaluation indicators. The calculation formula is as follows:

[0211]

[0212] In the formula, is the flood depth in meters; is the flow velocity in the flooded area in meters per second.

[0213] According to the evaluation index The flooded areas are divided into danger levels according to the threshold range, and a current flood danger level map for the entire river basin is prepared to provide an intuitive distribution of flood dangers in the entire river basin under the current flood status.

[0214] The real-time flood risk distribution point map obtained in S6-2 is superimposed on the satellite base map to obtain the measured data of each monitoring point in the basin, providing support for basin flood risk management.

[0215] This embodiment is based on a particle image recognition algorithm, and can directly use the naturally existing water splashes, foam or fallen leaves on the water surface as tracer particles for flow measurement. It can be applied to river scenes with different flow rates and flow states, and has a wider range of applications. At the same time, this embodiment can calculate the surface two-dimensional velocity field of the river where the video is shot, and can provide more comprehensive surface velocity field information of the river as a whole compared to the single-point or single-line velocity results obtained by traditional measurement methods.

[0216] Furthermore, this embodiment forms a real-time flood risk distribution point map by statistically analyzing the flow velocity and flow rate results of each flow measurement point. At the same time, the flow measurement results are combined with the two-dimensional hydrodynamic simulation model to deduce the flood inundation range and inundation depth under the current flow and water level state, so as to achieve an overall grasp of the flood risk distribution and flood danger of the entire river basin, and provide more data supplements and management basis for flood control in the river basin. Furthermore, this embodiment can be repeatedly called according to the needs of river basin flood risk management, and provide high-frequency real-time flood risk analysis results for the entire river basin.

[0217] like Figure 2 As shown, a specific embodiment of applying the present invention to conduct flood risk analysis on a key flood prevention village is as follows:

[0218] In a key flood prevention village, 12 river monitoring points (flow measurement points) have been deployed. It is set to collect a 10-second video every 10 minutes. The resolution of the cameras at the 12 flow measurement points is , the frame rate is 25 frames / second. Take the images in the video, keep one out of every five, and get a total of 50 images for each flow measurement point each time.

[0219] For each flow measurement point, 50 images are orthorectified using the four-point calibration method. Taking flow measurement point 1 as an example, first, four calibration points are selected within the video acquisition screen. Perform calibration, which includes the following:

[0220] The image is calibrated using the physical coordinates and pixel coordinates of the four calibration points. 1 Complete orthorectification to obtain , then the image Perform the same operation to obtain the orthorectified image .

[0221] The expressions of the physical coordinates and pixel coordinates of the four calibration points are as follows:

[0222]

[0223]

[0224]

[0225]

[0226] For this flow measurement point, the orthorectified top view image is calibrated according to the river range defined by the four calibration points. Perform rectangular cropping to obtain a A wide range of river flow images.

[0227] Based on orthorectified and cropped image , set horizontally in the image range Simulation grid points, vertical setting Simulation grid points .

[0228] Based on the particle image recognition algorithm, the cross-correlation calculation is performed on the images in the orthorectified image group in sequence. Create a contrasting area for the center pixels, and the next figure is represented by points The cross-correlation coefficients for the same contrast region created for the center are calculated with a time interval of 0.2 seconds between the two images. , select the corresponding The maximum displacement as water flows from Most likely displacement.

[0229] Each simulation grid point in the entire image is iteratively calculated to obtain a different maximum cross-correlation coefficient for each simulation grid point and select the corresponding displacement.

[0230] Therefore, for two consecutive images and The two-dimensional instantaneous velocity field of the surface between 2 , by calculating the Displacement Divide by time The calculation formula is as follows:

[0231]

[0232] So far we have obtained the image and 2, The two-dimensional instantaneous velocity field of the surface corresponding to the simulation grid points.

[0233] Through the image pair ( ),( ), ...( ) performs the above serial calculations to obtain 49 sets of two-dimensional surface instantaneous velocity field data.

[0234] Filter the two-dimensional instantaneous velocity field data of each image. Disassembled into ( ; ), that is, The velocity in two directions is defined according to the main direction of the flow velocity in the image (flowing from top to bottom). ; ) is expressed as follows:

[0235]

[0236]

[0237] Eliminate the result points that do not conform to the mainstream velocity direction; for the cross-correlation, set the minimum coefficient threshold, which can be 0.8, to filter out the velocity result points with a cross-correlation lower than 0.8, and finally obtain the two-dimensional instantaneous velocity field of each image. .

[0238] 49 sets of two-dimensional instantaneous velocity fields are averaged, which include the following:

[0239] According to the position of the simulation grid points, the average velocity of each simulation grid point is calculated by the following formula to obtain the two-dimensional average velocity field of the river surface corresponding to the video shooting period. The calculation formula is as follows:

[0240]

[0241] The river flow range in the image is , the physical coordinates of the measured river section are located at The water level information recognized by the camera water gauge during the video shooting period It is known that the underlying surface of the riverbed is mainly gravel.

[0242] 66 calculation points are evenly set along the straight line where the cross section is located , the interval between each calculation point is 0.5m. The river section has a bathymetric data every 1m. The bathymetric data of the section is linearly interpolated according to the interval of the calculation points to obtain the bathymetric data of each calculation point. At each calculation point Search for existing simulation grid points within the radius, average the two-dimensional instantaneous velocity results of the simulation grid points, assign them to the corresponding calculation points (center of the circle), and then perform linear interpolation on the calculation point data on the straight line where the section is located to obtain the new surface velocity results for each calculation point in the river section. .

[0243] According to the condition of the riverbed underlying surface, it is believed that the riverbed roughness is normal to smooth, so it can be set ,get , thereby converting the surface velocity of the river section into the average vertical velocity of the section.

[0244] The average flow velocity of each vertical line of the cross section, combined with the width of each vertical line sample strip and depth , to calculate the cross-sectional flow rate , and its calculation formula is as follows:

[0245]

[0246]

[0247] Among them, the cross-sectional water level , so that the cross-sectional flow of the river section in the basin during the shooting period can be calculated .

[0248] The maximum flow velocity, average flow velocity and cross-sectional flow rate of all flow measurement points in the period are counted. The maximum flow velocity of the 12 flow measurement points is calculated by calculating the maximum value of the average flow velocity of all cross-sectional vertical lines of the flow measurement points. The average flow velocity is obtained by calculating the average of the average flow velocity of all cross-sections at the flow measurement point. get.

[0249] Combined with the flood risk classification standard of the basin, the maximum flow velocity, average flow velocity and cross-sectional flow are divided into different risk levels, and the assessment threshold is set. For most basins, the flow velocity threshold is , the water depth threshold is , the end flow threshold is .

[0250] Taking flow measurement point 1 as an example, the maximum flow rate , which is less than the flow rate threshold , and the maximum water depth of the river section is less than the water depth threshold ; mean flow rate , which is less than the flow rate threshold ; Cross-sectional flow , which is less than the end flow threshold , and therefore assessed as no flood risk.

[0251] By integrating the risk classification results and flow measurement results of flow measurement points in the entire river basin, a real-time flood risk distribution point map of the entire river basin is obtained, which is visualized on the digital twin platform. The result data of each flow measurement point is uploaded every 10 minutes to provide data support for basin flood control management.

[0252] At the same time, according to the monitoring range of the watershed, an overall unstructured two-dimensional grid is built, with a grid density of 1-5m at the embankment, 5-10m in the river channel, 5-30m in key flood-prone villages and towns, and 30-100m in the mountainous areas at the edge of the range. The elevation data of the watershed with an accuracy of 5m and the elevation data of the river section are interpolated onto the grid, and the grids near the embankment and key flood-prone villages and towns are adjusted to accurately express the topographic changes of the embankment and villages and towns.

[0253] Furthermore, since there are two tributaries in the basin near a key flood control village, they converge in the middle of the basin to form the main stream. The upstream flow measurement points corresponding to the two tributaries are No. 1 and No. 2, respectively, and the downstream flow measurement point corresponding to the main stream is No. 12. Therefore, the cross-sectional flow data calculated by flow measurement points No. 1 and No. 2 are used as the boundary flow input of the two upstream tributaries of the model, and the water level data obtained are used as the water level boundary conditions of the two upstream tributaries of the model; the water level information obtained by flow measurement point No. 12 is used as the water level boundary of the downstream model. The water level information in the river channel obtained from the 12 flow measurement points is assigned to the model as the initial water level condition in the river channel through linear interpolation.

[0254] Different roughness coefficients are set for the model according to the conditions of the riverbed underlying surface and different areas on the shore, such as woodlands, farmlands, and towns. Then, a two-dimensional hydrodynamic simulation calculation is performed with a simulation time of 1 hour. Based on the computing resources and the size of the basin, the simulation calculation time for the flood inundation situation in the basin is set to 3 minutes, so as to obtain the calculation results of the basin.

[0255] The last frame of the calculation result of the basin is intercepted to extract the water depth field data. According to the location coordinates of the bank line and the top of the river bank in the basin, the result data in the river channel is eliminated and the water depth field data on the bank is retained.

[0256] The grid nodes with water depth greater than or equal to 0.001m in the two-dimensional water depth results were counted as submerged nodes. The submerged nodes with water depth equal to 0.001m were connected, and contour lines were drawn as the submerged range, resulting in three main submerged areas in the basin and several other submerged areas.

[0257] According to the designated flooding range, the area of ​​the grid within the range is counted, and the total flooding area of ​​the basin in the current state is 0.28 square kilometers. The flooding areas of the three main flooding areas are 0.01 square kilometers, 0.07 square kilometers and 0.15 square kilometers respectively. The water depth results of the flooded areas are divided into 7 levels according to data such as 0.001-0.2m, 0.2-0.4m, 0.4-0.6m, 0.6-0.8m, 0.8-1.0m, 1.0-1.2m and greater than 1.2m, and contour lines are drawn.

[0258] The velocity field data of the last frame of the calculation results of the basin is intercepted, and the results are divided into the river channel part and the shore part according to the location coordinates of the river bank and the top of the river bank in the basin. For the flooded area on the shore, the velocity results are divided into 8 levels according to 0.0-0.4m / s, 0.4-0.8m / s, 0.8-1.2m / s, 1.2-1.6m / s, 1.6-2.0m / s, 2.0-2.4m / s, 2.4-2.8m / s and greater than 2.8m / s, and contour lines are drawn. Contour lines are also drawn for the velocity in the river channel according to this division method, and according to the flood risk classification standard of the basin, the river section with a velocity exceeding 0.5m / s is additionally delineated by contour lines to produce a velocity risk map.

[0259] According to the inundation range of the designated river basin, the water depths correspond from light red to dark red from low to high, and a flood inundation map of the current state of the river basin is produced, providing areas in the river basin that may be at risk of inundation under the current flood state and their possible degree of inundation.

[0260] According to the flood risk level classification standard of the basin, the flood depth and flow rate are used as evaluation indicators. The calculation formula is as follows:

[0261]

[0262] In the formula, is the flooding depth in meters (m); is the flow velocity in the flooded area in meters per second (m / s).

[0263] The classification standards for flood risk levels in river basins are as follows:

[0264] when When , the danger level is low;

[0265] when When the risk level is medium;

[0266] when , the danger level is high.

[0267] The inundated areas of the entire river basin are divided into three levels: low, medium and high, so that the current flood hazard level map of the entire river basin can be obtained, thereby providing the distribution of flood hazards in the entire river basin under the current flood status.

[0268] A system embodiment using the method of the present invention:

[0269] A basin flood risk assessment system based on particle recognition algorithm, comprising:

[0270] one or more processors;

[0271] A storage device for storing one or more programs;

[0272] When the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned basin flood risk assessment method based on particle recognition algorithm and form a digital twin system that can be visualized.

[0273] In this embodiment, the flood inundation map, flow velocity risk map and flood hazard level map under the current status of the entire river basin are superimposed on the satellite base map to obtain the distribution of flood hazards in the entire river basin, and visualized on the digital twin system to provide river basin managers with a clear and intuitive overall situation of river basin flood risks.

[0274] The digital twin system of the present invention cyclically collects flow measuring point videos and water level data at intervals of 10 minutes, performs calculations and analysis, obtains real-time flow velocity and flow data for each flow measuring point, and uploads the data to the system, thereby being able to deduce the flood inundation situation in the current state, thereby forming a visualization effect, and providing real-time dynamic data and analysis results to support flood control in the basin.

[0275] A computer medium embodiment of the method of the present invention is applied:

[0276] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned basin flood risk assessment method based on a particle recognition algorithm.

[0277] Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, and computer program products. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, optical storage, etc.) containing computer-usable program codes.

[0278] The present application is described by flowcharts or / and block diagrams of the methods, devices (systems), and computer program products of the embodiments of the present application. It should be understood that each process or / and block in the flowchart or / and block diagram and the combination of the processes or / and blocks in the flowchart or / and block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart or / and block diagram. Figure 1 Process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0279] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 Process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0280] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 Process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0281] The model in this application is an object that objectively describes the morphological structure with the help of physical or virtual representation. The object is not equal to the object and is not limited to physical and virtual. It can be a data processing function, software program, processing mode, usage method, operation method, workflow, application process, electronic hardware, circuit module, processing system, system imitation or simulation object.

[0282] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field can still modify or replace the specific implementation methods of the present invention with equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be included in the scope of protection of the claims of the present invention.

Claims

1. A basin flood risk assessment method based on particle recognition algorithm, characterized by: Includes the following: Through the pre-built watershed spatial discrete model, a watershed and surrounding terrain are modeled to obtain a watershed simulation object; Using the pre-built velocity calculation model, one or more groups of water flow images of the watershed simulation object are obtained, and based on the particle image recognition algorithm, one or more groups of water flow images are serially analyzed to calculate the velocity field information of the watershed; Using the pre-built flow simulation model, based on the basin section location and section elevation data of the basin simulation object, the flow velocity field information of the basin is calculated to obtain one or more section flows; According to the pre-built two-dimensional hydrodynamic model, based on the cross-sectional flow and water level data of the basin simulation object, the flood inundation situation of the basin simulation object is simulated to obtain the flood change simulation result; the method is as follows: Obtain the hydrodynamic data of the basin simulation object, including water depth, water surface elevation, sampling period, average water velocity value, basin bottom friction coefficient, surface wind stress, Coriolis force, atmospheric pressure and turbulent viscosity diffusion coefficient; Based on the hydrodynamic data, the continuity equation and momentum equation are established to form a two-dimensional hydrodynamic equation group; The cross-sectional flow and water level data of the basin simulation object are input into the two-dimensional hydrodynamic equations as initial conditions and boundaries, and the basin simulation calculation is performed to obtain the two-dimensional water depth field, water level field and flow velocity field of the basin; Based on the two-dimensional water depth field and water level field of the basin, combined with the surface elevation data, the water depth data belonging to the water area is eliminated to obtain the flooded water depth data; According to the flooded water depth data, grid nodes with water depth greater than or equal to 0.001 meters are counted as flooded nodes, and the flooded nodes with water depth equal to 0.001 meters are connected to obtain contour lines; The area surrounded by the contour lines is regarded as the flooded area; and calculate the area of ​​the flooded region as the flooded area; Based on the velocity field, the velocity field in the water area and the velocity field on land are distinguished according to the water shoreline, and the velocity in the flooded area and the watershed section exceeding the velocity safety range are obtained; Based on the inundation area, inundation area, inundation depth data, flow velocity in the inundation area and the watershed section exceeding the flow velocity safety range, the flood inundation situation of the watershed simulation object is determined to obtain the flood change deduction results; Through the pre-built flood risk assessment model, the flood change simulation results are evaluated, the flood risk distribution of the basin simulation object is obtained, and the basin flood risk assessment based on the particle recognition algorithm is completed.

2. A basin flood risk assessment method based on particle recognition algorithm as claimed in claim 1, characterized in that: The method of modeling a watershed and its surrounding terrain through the pre-constructed watershed spatial discrete model to obtain the watershed simulation object is as follows: Get the scope of a watershed to be evaluated; According to the scope to be assessed, obtain elevation data, basin section location and water level information of several locations; Based on the elevation data and the location of the watershed section, a two-dimensional grid divided in an unstructured manner is constructed; Based on the terrain characteristics, the two-dimensional grid is locally refined to represent the topographical changes of the embankment and / or villages and towns; Perform linear interpolation on the water level information to form water level data, and assign it to the two-dimensional grid as the water level condition; According to the conditions of the underlying surface of the watershed and different areas on the shore, several roughness coefficients are set; The water level condition, roughness coefficient and two-dimensional grid are coupled to obtain the watershed simulation object.

3. A basin flood risk assessment method based on particle recognition algorithm as claimed in claim 1, characterized in that: Using the pre-built velocity calculation model, one or more groups of water flow images of the watershed simulation object are obtained, and based on the particle image recognition algorithm, one or more groups of water flow images are serially analyzed to calculate the watershed velocity field information as follows: Constructing a water flow image acquisition unit for acquiring one or more groups of water flow images of a certain watershed; Constructing a velocity analysis unit for serially analyzing one or more groups of water flow images and calculating the instantaneous velocity field data of water flow at a number of simulation grid points; Decompose the instantaneous velocity field data of water flow into lateral velocity and longitudinal velocity; According to the value range limited by the flow velocity direction, the lateral velocity and the longitudinal velocity are processed, and the instantaneous velocity field data of the water flow that does not conform to the flow velocity direction are eliminated to obtain the first instantaneous velocity field data of the water flow; According to the minimum coefficient threshold of the cross-correlation, the first water flow instantaneous velocity field data is filtered to obtain the second water flow instantaneous velocity field data; According to the position of the simulation grid points, the instantaneous velocity field data of the second water flow are averaged to calculate the two-dimensional average velocity field of the basin surface corresponding to the video shooting period, that is, the basin flow velocity field information.

4. A basin flood risk assessment method based on particle recognition algorithm as claimed in claim 3, characterized in that: The method of constructing the water flow image acquisition unit is as follows: Based on the statistical frequency of hydrological data in the basin and the demand for points, determine the required flow measurement points in the basin and count the number of flow measurement points; Set the video collection duration information based on the frequency of hydrological data collection; According to the video acquisition duration information, the video acquisition data of the flow measurement point is obtained, and the resolution and frame rate of the camera are obtained; Perform image acquisition on the video acquisition data, retain one image for every n images, and number the images in chronological order to form an initial water flow image group; Based on the distance between the markers on both sides of the watershed in the water flow image, the images in the initial water flow image group are orthorectified to restore the image's scale and spatial geographic coordinate information in reality, and obtain one or more groups of orthorectified orthographic image groups; The top view image group is used as the water flow image of a certain river basin to complete the construction of the water flow image acquisition unit; The method of constructing a speed analysis unit is as follows: According to the required calculation accuracy of the two-dimensional velocity field and the width of a certain watershed, the required number of simulation grid points is calculated; The number of simulation grid points includes the number of horizontal grid points and the number of vertical grid points; According to the number of horizontal grid points and the number of vertical grid points, the orthorectified water flow image is grid-divided to obtain a number of simulation grid points; Based on the simulation grid points on two continuous water flow images, a correlation calculation unit is constructed to obtain the surface two-dimensional instantaneous velocity field between the two continuous water flow images; Utilizing the correlation calculation unit, serial correlation calculation is performed on all water flow images to obtain the two-dimensional surface instantaneous velocity field between the water flow images, namely, the instantaneous velocity field data of the water flow, thus completing the construction of the velocity analysis unit.

5. A basin flood risk assessment method based on particle recognition algorithm as claimed in claim 4, characterized in that: Based on the distance between the landmarks on both sides of the watershed in the water flow image, the method for orthorectifying the images in the initial water flow image group is as follows: When the image is shot at a top-down angle, the scale calibration method is used for orthorectification. Includes the following: Obtain two mutually perpendicular landmarks in the initial water flow image group; For two mutually perpendicular markers, connect them to obtain the physical distance and pixel distance between the two markers; Determine the horizontal and vertical proportional relationship between physical distance and pixel distance; All images are scaled in proportion to obtain one or more groups of orthographic image groups after orthorectification; When the image is taken at an oblique viewing angle, the four-point calibration method is used for orthorectification, which includes the following: Select four markers in the initial water flow image group, any three of which are not collinear; Calculate the physical distances between four landmarks; According to the physical distance, the relative physical coordinates of the four markers are determined and used as the coordinates of the calibration points; Based on the calibration point coordinates, the conversion coefficient between the image physical coordinates and the pixel coordinates is obtained; Apply the conversion coefficient to each pixel in the image to complete the orthorectification of the image; The orthorectified image is cropped in a rectangular shape to retain the area where the flow velocity and flow rate need to be calculated in the water flow image, so as to obtain one or more groups of orthographic image groups; Or / and, based on the simulation grid points on two continuous water flow images, a correlation calculation unit is constructed to obtain the surface two-dimensional instantaneous velocity field between the two continuous water flow images as follows: Create a comparison area 1 with a simulation grid point in a water flow image as the center; According to the serial number of a water flow image, adjacent water flow images are obtained; Create a second comparison area centered on the simulated grid points of the adjacent water flow images; Obtain pixel values ​​and pixel intensity distributions of contrast area 1 and contrast area 2 respectively, and set a search range; Based on the search range, pixel value and pixel intensity distribution, the correlation between the comparison area 1 and the comparison area 2 is calculated to obtain a correlation coefficient; According to the correlation coefficient, the displacement of the simulation grid points within a certain time period is calculated; According to the displacement and the corresponding time period, the two-dimensional instantaneous velocity field of the surface between two consecutive water flow images is obtained.

6. A method for assessing river basin flood risk based on a particle recognition algorithm as claimed in claim 5, characterized in that: Using the pre-built flow simulation model, based on the basin section location and section elevation data of the basin simulation object, the flow velocity field information of the basin is calculated to obtain the cross-sectional flow as follows: Based on the location and elevation data of the watershed section and the velocity field information of the watershed, the surface velocity on a watershed section is obtained; The conversion coefficient is calculated based on the roughness of the basin where the watershed section is located, the water velocity at a certain depth, and the ratio of a certain depth to the water depth of the section; According to the conversion coefficient, the surface velocity of the watershed section is converted into the average velocity of the cross-section perpendicular to the cross-section; The cross-sectional flow rate is calculated by the average flow velocity of the cross-sectional vertical line and the cross-sectional water depth, combined with the width and depth of each cross-sectional vertical line.

7. A method for assessing river basin flood risk based on particle recognition algorithm as claimed in claim 6, characterized in that: Based on the basin section location and section elevation data, and according to the basin velocity field information, the method for obtaining the surface velocity on the basin section is as follows: Based on the location of the watershed section, a simulation straight line where the section is located is constructed; On the simulation line, multiple calculation points are evenly set; According to the cross-section elevation data, the interval distance value between the calculation points is obtained; Take the calculation point as the center of the watershed circle, set the interval distance value as the search radius, and get the search area; Searching for existing simulation grid points in the search area, obtaining a number of simulation grid points and corresponding flow velocity field information of the watershed; Averaging the velocity field information of several watersheds to obtain the average velocity field information; The average velocity field information is used as the surface velocity of the calculation point and assigned to the corresponding calculation point, that is, the center of the basin; Based on the center of the basin and the corresponding average flow velocity field information, linear interpolation is performed on the calculation points on the simulation line to obtain the surface velocity of each calculation point on the basin section.

8. The method for assessing river basin flood risk based on particle recognition algorithm according to claim 1, characterized in that: The method of conducting risk assessment on the flood change simulation results through the pre-built flood risk assessment model and obtaining the flood risk distribution of the basin simulation object is as follows: According to the flood range and flood area, a flood inundation map of the basin simulation object is produced to obtain the areas with flooding risks under the current flood state; According to the area where there is a risk of flooding, obtain the flooding depth and flow rate of the area; Multiply the flood depth and flow velocity to obtain the flood risk assessment index; According to the inundation risk assessment index and the basin flood hazard level classification standard, the inundated area is divided into hazard levels, and the current flood hazard level map is drawn to obtain the flood risk distribution of the basin simulation object under the current flood state.

9. A basin flood risk assessment system based on particle recognition algorithm, characterized by: It includes: one or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement a watershed flood risk assessment method based on a particle recognition algorithm as described in any one of claims 1 to 8.

Images