Automatic crack flow identification method and system based on flow field data

Abstract

The invention discloses a split flow automatic identification method and system based on flow field data, and the method comprises the steps: building a grid at a target location, obtaining the flow velocity and flow direction of each grid point, extracting connected objects, and calculating the area and the minimum bounding box of each connected object; the communication objects with the areas larger than a first threshold value and the included angles between the long axes of the communication objects and the offshore direction of the target site smaller than a second threshold value are reserved, and the communication objects are judged as possible flow splitting units; the possible flow splitting units at multiple moments are tracked until all the possible flow splitting units are marked with serial numbers; the label split flow of which the duration time is greater than a third threshold value is reserved; and superposing the split flow of each serial number on the flow field graph to draw a minimum bounding box, and performing manual verification to obtain a split flow identification result. The method can identify two types of split flows of groove split flow and transient split flow, can be suitable for flow field data of different sources, and is used for early warning and protection of coastal split flow.

Description

Technical Field

[0001] The present invention relates to the technical field of rip flow identification, in particular to a method and system for automatically identifying rip flow based on flow field data. Background Art

[0002] Rip currents are one of the most important marine hazards in coastal tourist areas. They are narrow, fast, short-lived and changeable, and can draw in unsuspecting beachgoers and drag them into deep waters, leading to drowning accidents. Rip currents are prone to occur on beach terrains with sandbar grooves, so current meters can be deployed at single points according to the seabed terrain for monitoring and early warning. Studies have found that a type of transient rip current may also occur on beaches without undulating seabed terrain. This type of rip current is different from the fixed-position groove rip current. When it occurs, it shows the characteristics of position migration and has a short duration. Therefore, its monitoring requires obtaining the full-field flow field information of the beach.

[0003] However, the identification of rip currents from flow field data is still mainly done through manual visual interpretation, and there is a lack of automatic identification methods that can systematically extract information such as rip current characteristics from flow field data. Manual analysis is mainly based on the existing description of rip currents and some field observations of rip current velocity. Rip currents usually occur in the wave breaking zone and can extend beyond the wave breaking zone. They have a large velocity, usually 0.5-2m / s, a narrow width, and present an offshore flow direction opposite to its two sides. They are often accompanied by converging coastal currents to form a circulation system. However, the above description is more applicable to fixed-position groove rip currents, and does not fully cover the characteristics of transient rip currents, which are not fixed in position and have a short duration. In addition, since the formation of transient rip currents is often accompanied by the generation of vortices in the wave breaking zone, some scholars use the method of calculating vorticity and vortex detection to indirectly infer the occurrence of rip currents, but this method is not a direct determination of rip currents, and some vortices that do not have offshore current characteristics will be misjudged as rip currents, and it is not applicable to the identification of groove rip currents at the same time. It can be seen that there are great differences in the spatial location and duration of groove rift flow and transient rift flow. Currently, there is a lack of unified methods to systematically and automatically identify them. Summary of the invention

[0004] Purpose of the invention: The purpose of the present invention is to provide a method and system for systematically extracting rip flow characteristics from flow field data and automatically identifying and tracking groove rip flow and transient rip flow.

[0005] Technical solution: The method for automatic identification of rip current based on flow field data of the present invention comprises the following steps:

[0006] (1) Establish grids along the offshore and coastal directions at the target location and obtain the flow velocity and direction of each grid point at each moment;

[0007] (2) Search for possible rip current units at each moment. For grid points where all flow velocities and flow directions are within the threshold range, extract connected objects and calculate the area and minimum bounding box of each connected object; retain connected objects with an area greater than the first threshold and an angle between the major axis of the connected object and the offshore direction of the target location less than the second threshold, and determine them as possible rip current units;

[0008] (3) Track possible rip current units at multiple moments. For each possible rip current unit at each moment, the other possible rip current units are its candidate units. Select the candidate unit that is closest to the possible rip current unit in the offshore direction, and mark the two as rip currents with the same serial number; repeat step (3) until all possible rip current units have been marked with serial numbers;

[0009] (4) Superimpose and draw the minimum bounding box of each serial number of rip current on the flow field diagram, and obtain the rip current recognition result through manual verification.

[0010] Further, in step (1), obtain the flow field data of the target location and calculate the time-averaged flow velocity field. Establish a grid along the offshore direction and the alongshore direction, and use the linear interpolation method to obtain the offshore flow velocity u and the alongshore flow velocity v of each grid point, and calculate the flow velocity and flow direction

[0011] Further, in step (2), extracting connected objects and calculating the area and minimum bounding box of each connected object includes:

[0012] If the flow velocity s and flow direction θ of the grid point satisfy s≥U and θ<α, then generate a mask with a mask value of 1 for this grid point, otherwise generate a mask with a mask value of 0; U is the rip current velocity threshold, and α is the rip current flow direction threshold;

[0013] Use 8-connected neighborhood connectivity to extract connected objects in the mask, calculate the area of each connected object, and the minimum bounding box of the connected object.

[0014] Further, in step (2), before retaining connected objects with an area greater than the first threshold and an angle between the major axis of the connected object and the offshore direction of the target location less than the second threshold, it also includes:

[0015] Judge whether the area a of the connected object satisfies a≥B 2 , if not, delete the connected object; B is the geometric threshold;

[0016] Calculate the minimum distance d between the minimum bounding boxes of two connected objects, merge the two connected objects that satisfy d<B, and update the area and minimum bounding box of the connected object;

[0017] The connected objects whose side close to the shoreline is nearly parallel to the shoreline are deleted, and the area and minimum bounding box of the connected objects are updated.

[0018] Further, the determining whether the connected objects include a connected object whose boundary is parallel to the shoreline, and if so, deleting the connected object, and updating the area and minimum bounding box of the connected object includes:

[0019] (5.1) For the connected object S, its minimum bounding box is divided into two equal parts along the coastline to obtain two connected objects S1 and S2. S1 is the connected object close to the coastline.

[0020] (5.2) Calculate the angle between the long axis of S1 and S2 and the offshore direction respectively. If the angle is greater than α, take the long axis length of the connected area as its characteristic width along the shore. Otherwise, take the short axis length to obtain the characteristic widths w1 and w2 of S1 and S2. α is the rip current direction threshold.

[0021] (5.3) Calculate the median width w3 of S2 along the coastal direction;

[0022] (5.4) If the judgment condition max(w1,w2)>2B and w1>1.5*w3 is met, then the minimum bounding box of S is moved one grid unit offshore from the boundary of the coastal side to obtain a new minimum bounding box, and steps (5.1) to (5.3) are repeated until the judgment condition is not met, or w1 or w2 is 0, and the updated minimum bounding box is obtained;

[0023] (5.5) Use the updated minimum bounding box to extract the updated connected object S'.

[0024] Furthermore, in step (2), the connected objects whose reserved area is greater than the first threshold and whose angle between the long axis of the connected object and the offshore direction of the target location is less than the second threshold include:

[0025] Delete the connected object whose area a does not satisfy a≥B 2 connected objects, B is the geometric threshold;

[0026] Delete the connected objects whose angle γ between the major axis of the minimum bounding box of the connected object and the offshore direction does not satisfy γ<1.33α, where α is the rip current direction threshold.

[0027] Furthermore, in step (3), before selecting a candidate unit that is closest to the possible rip current unit in the offshore direction and marking both of them as rip currents with the same serial number, the method further includes:

[0028] For each possible rift unit at each moment, it is moved by B sinβ / sin45° in the coastal direction and extended by B cosβ / cos 45° in the offshore direction. If the overlap area between the minimum bounding box of the candidate rift unit and the moved rift unit is less than 60%, the candidate unit is deleted; β is the average flow direction on the offshore boundary of the possible rift unit, and B is the geometric threshold.

[0029] Further, in step (3), a candidate unit that is closest to the possible rip current unit in the offshore direction is selected, and the rip currents that mark the two as having the same serial number include:

[0030] For each possible rip current unit at each moment, if its displacement to the candidate unit is not positive in the offshore direction, the candidate unit is deleted;

[0031] Among the remaining candidate units, the distance between the center point of the minimum external frame of the possible split flow unit and the center point of the minimum external frame of the candidate unit is calculated, the candidate unit with the shortest distance is selected, and the possible split flow unit and the candidate unit are marked as split flows with the same serial number.

[0032] Furthermore, the duration of each labeled rip flow is calculated, and the rip flows with a duration less than a third threshold are deleted.

[0033] The automatic rip flow identification system based on flow field data of the present invention comprises:

[0034] A data acquisition unit is used to establish a grid at a target location along an offshore direction and a coastal direction, and obtain a flow velocity and a flow direction at each grid point at each moment;

[0035] A possible rift unit search unit is used to search for possible rift units at each moment. For all grid points whose flow velocity and flow direction are within the threshold range, connected objects are extracted and the area and minimum bounding box of each connected object are calculated; connected objects whose area is greater than a first threshold and whose angle between the long axis of the connected object and the offshore direction of the target location is less than a second threshold are retained and determined as possible rift units;

[0036] The possible rift unit tracking unit is used to track possible rift units at multiple moments. For each possible rift unit at each moment, the remaining possible rift units are its candidate units, and a candidate unit closest to the possible rift unit in the offshore direction is selected, and both are marked as rifts with the same serial number; the possible rift unit tracking unit is repeatedly executed until all possible rift units have been marked with serial numbers;

[0037] The rip flow identification and verification unit is used to superimpose each rip flow number on the flow field diagram to draw its minimum external bounding box, and obtain the rip flow identification result through manual verification.

[0038] Beneficial effects: Compared with the prior art, the advantages of the present invention are as follows: The present invention is proposed to address the current limitations of rip current identification that requires manual interpretation, and the current situation that the temporal and spatial characteristics of beach rip currents are variable and there is a lack of a systematic method for identifying and judging them. The present invention analyzes the flow velocity and direction of each grid point from the flow field data, thereby identifying possible rip current units and tracking their direction, marking the rip currents, and being able to identify two types of rip currents: groove rip currents and transient rip currents; it can be applied to flow field data from different sources, and can be widely used in sea areas with high incidence of rip currents such as leisure and tourism beaches, and can provide technical support for early warning and protection of seaside rip currents, and improving beach safety in coastal tourist areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A flow chart of a method for automatically identifying rip current according to an embodiment of the present invention;

[0040] Figure 2 is a schematic diagram of connected objects in an embodiment of the present invention;

[0041] Figure 3 It is a time-averaged flow field diagram and a schematic diagram of the minimum external frame of a rip flow unit in Example 1 in an embodiment of the present invention;

[0042] Figure 4 It is a time-averaged flow field diagram and a schematic diagram of the minimum external frame of a rip flow unit in Example 2 in an embodiment of the present invention;

[0043] Figure 5 It is a time-averaged flow field diagram and a schematic diagram of the minimum external frame of a rip flow unit in Example 3 in an embodiment of the present invention;

[0044] Figure 6 It is a time-averaged flow field diagram and a schematic diagram of the minimum external frame of the rip flow unit of Example 4 in an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0046] like Figure 1 As shown, the method for automatically identifying rip flow based on flow field data includes the following steps.

[0047] Step 1: Collect flow field data and perform preprocessing.

[0048] Step 1.1: average the collected flow field data and output the average flow velocity field per minute.

[0049] Step 1.2: A two-dimensional square grid is established along the offshore direction x and the onshore direction y, and the flow field data is linearly interpolated to obtain the offshore current velocity u and the onshore current velocity v at each grid point.

[0050] Step 1.3, calculate the flow velocity at each grid point and flow direction

[0051] Step 2: Search for possible rift flow units in the flow field at each moment.

[0052] Step 2.1, set the rip flow velocity threshold U, rip flow direction threshold α, and geometric threshold B.

[0053] Step 2.2, determine whether the velocity magnitude and direction of each grid point meet the requirements of s≥U and θ<α, and generate a mask based on the determination result. If the above conditions are met, the mask value is 1, otherwise it is 0.

[0054] Step 2.3, use 8-connected neighborhood connectivity to extract connected objects in the mask, calculate the area a of each connected object and the position of the minimum bounding box containing the area.

[0055] Step 2.4: Determine whether the area of ​​the connected object meets a≥B 2 If it does not meet the requirements, the object will be deleted from the object pool.

[0056] Step 2.5, calculate the minimum distance between the minimum bounding boxes of the connected objects in the object pool. If the minimum distance between the two boxes is less than B, the two connected objects are considered to be the same connected object, and the positions of their corresponding minimum bounding boxes are updated, and the object pool is updated.

[0057] Step 2.6: For all connected objects in the object pool, delete the connected objects that are close to the shoreline and are nearly parallel to the shoreline, such as Figure 2 As shown, the specific method is as follows:

[0058] 1) For an object S, divide its minimum bounding box into two halves along the y-axis direction, and use the method in step 2.3 to obtain two connected objects S1 and S2, and designate S1 as the connected object on the coast side;

[0059] 2) Calculate the angle γ between the long axis of S1 and S2 and the x-axis respectively. If the angle γ>threshold α, take the long axis length of the connected object as its coastal characteristic width, otherwise take its short axis length, and obtain the characteristic widths w1 and w2 for the two connected objects S1 and S2 respectively;

[0060] 3) For the connected object S2, calculate the median width w3 along the coast;

[0061] 4) Make the following judgment: max(w1,w2)>2B and w1>1.5*w3. If so, move the minimum bounding box of S close to the shoreline side boundary toward the offshore direction by one grid unit, and repeat the above steps 1) to 4) after obtaining the new minimum bounding box;

[0062] 5) If the condition in step 4) is not met, or w1 or w2 is 0, the update is stopped. At this time, the method in step 2.3 is used to extract the connected object from the minimum bounding box updated in the previous round, which is the updated connected object S'.

[0063] Step 2.7: Determine whether the area of ​​the connected object updated in step 2.6 satisfies a ≥ B 2 If it does not meet the requirements, the connected object will be deleted from the object pool.

[0064] Step 2.8, determine whether the angle between the long axis of the connected object in step 2.7 and the x-axis direction is less than 1.33α, if not, delete the object from the object pool.

[0065] Step 2.9, treat each individual connected object in the object pool in step 2.8 as a possible rift unit, record its time, minimum bounding box position, and the average flow direction β on the offshore boundary of the connected object (the average flow direction of all grid points on the offshore boundary of the connected object).

[0066] Step 3: Based on the results of step 2, track the movement of possible rift flow units at multiple times.

[0067] Step 3.1, set the rip flow time threshold T.

[0068] Step 3.2, for each possible rip flow unit at each moment obtained in step 2, follow the following steps in order of time:

[0069] 1) For all possible rip flow units at each moment, determine whether they have been numbered one by one. If not, mark them with a new serial number;

[0070] 2) For all possible rip flow units described in 1), perform the following steps one by one:

[0071] 2.1) For a possible split flow unit R_0, all k possible split flow units that are not labeled between t_0+1 and t_0+T are listed as candidate units R_n (n=1…k);

[0072] 2.2) Move the minimum outer frame of the R_0 unit by B sinβ / sin45° in the y-axis direction and extend it by B cosβ / cos 45° in the x-axis direction. If the area where the minimum outer frame of the candidate unit overlaps with it is not greater than 60%, the candidate unit is deleted from the candidate pool;

[0073] 2.3) Calculate the displacement from R_0 to each candidate unit. If the displacement is not positive in the x-axis direction, delete the candidate unit from the candidate pool.

[0074] 2.4) If the number of remaining candidate units is not 0, the candidate unit with the shortest distance between the center point of the smallest external frame and the center point of the smallest external frame of R_0 is determined to be the same rift unit as R_0 at time t_0+1, and the two possible rift units are marked as rifts with the same serial number, and the remaining candidate units are deleted from the candidate pool;

[0075] 2.5) The candidate units deleted in the above steps 2.2) to 2.4) remain unlabeled and enter the next round of tracking.

[0076] Step 3.3, repeat the above tracking process until the data ends.

[0077] Step 3.4, check whether the duration of the rip flow labeled in step 3.3 is less than the time threshold T, if so, delete it from the rip flow pool.

[0078] Step 3.5, record the characteristic information such as the starting time, maximum flow velocity, length, width, duration and the minimum external frame position of the possible rip flow unit at each moment of the rip flow labeled in step 3.4.

[0079] Step 4, verify the rip current determined in steps 2 to 3 through manual interpretation and comparison.

[0080] Step 4.1: For the flow field data selected for debugging, draw the minute-average flow field diagram.

[0081] Step 4.2, superimpose on the flow field diagram the minimum outer bounding box line of the rip flow unit that appears at each moment within the starting time of the rip flow determined by steps 2 to 3.

[0082] Step 4.3: Check visually whether there are any unidentified obvious rip currents and whether the flow field within the minimum bounding box of the identified rip currents meets the definition of rip currents.

[0083] Step 4.4, adjust the geometric threshold B according to the comparison result of step 4.3 until the identification result reasonably conforms to the rip flow definition.

[0084] Furthermore, if this method is used for the first time at a certain location, it is necessary to first select part of the flow field data and use steps 2 and 3 to determine the rip flow, then compare it with the manual interpretation, check whether the automatically identified rip flow meets the definition, and adjust the threshold parameters in steps 2 and 3. After determining the parameters suitable for this location, use the adjusted threshold parameters to determine the rip flow of the remaining flow field data.

[0085] The method of the present invention is verified by specific experiments below.

[0086] (1) Obtain flow field data. In this embodiment, nearshore flow field data simulated based on the Bousinnesq wave model FUNWAVE-TVD is used. The model setting method and general parameters are all described in the model manual, so they are not repeated here. The simulation area is a rectangular area of ​​500m along the coast and 512m across the coast. A wave maker based on the TMA spectrum is used, and the wave maker is located 400m offshore. The examples simulated in this embodiment include groove rip currents under sand bar terrain and transient rip currents on a straight and uniform beach, as shown in Table 1. During the model operation, the direct output is the average flow rate per minute.

[0087] Table 1 Wave-making conditions and terrain settings for the simulation of nearshore flow data

[0088]

[0089] A two-dimensional square grid with a horizontal resolution of 1 meter was established along the offshore direction x and the alongshore direction, and the flow velocity and direction at each grid point were calculated.

[0090] (2) For the uniform terrain examples 1 and 2 in Table 1, the rip current velocity threshold U = 0.5 m / s, the direction threshold α = 45°, and the geometric threshold B = 10 m; for the groove terrain examples 3 and 4 in Table 1, the rip current velocity threshold U = 0.5 m / s, the direction threshold α = 60°, and the geometric threshold B = 15 m. Identify all possible rip current units, record and save their time, minimum outer frame position, and the average flow direction β on their offshore boundary.

[0091] (3) For Examples 1-4 in Table 1, the rip time threshold T = 2 minutes is set, and the possible rip units are tracked and numbered, and characteristic information such as the starting time, maximum flow velocity, length, width, duration, and the minimum outer frame position of the possible rip units at each moment are recorded.

[0092] (4) Draw the time-averaged flow field diagram at each output moment of the simulated flow field in Table 1, and superimpose the minimum bounding box of the corresponding possible rift flow unit in the rift flow pool on the time-averaged flow field diagram. The time-averaged flow field diagrams and the minimum bounding boxes of the possible rift flow units of Examples 1 to 4 are shown in Figure 1. Figures 3 to 6 shown.

[0093] Check the result graph to see if there are any cases where rip currents are not identified or the flow field within the minimum outer frame does not meet the definition of rip currents. Figures 3 to 6 The rift flow unit marked by the smallest outer frame line is consistent with the manual interpretation, among which, Figure 2 , 3 5 are transient rift currents. It can be seen that the rift current position migrates along the coast and offshore at different times, and the rift current gradually sheds off the vortex moving offshore, which is consistent with its generation and evolution mechanism. Figure 4 It is worth noting that there are multiple rifts in the flow field at the same time, but their positions do not change at different times. This shows that this method can identify two types of rifts.

[0094] The automatic rip flow identification system based on flow field data of the present invention comprises:

[0095] A data acquisition unit is used to establish a grid at a target location along an offshore direction and a coastal direction, and obtain a flow velocity and a flow direction at each grid point at each moment;

[0096] A possible rift unit search unit is used to search for possible rift units at each moment. For all grid points whose flow velocity and flow direction are within the threshold range, connected objects are extracted and the area and minimum bounding box of each connected object are calculated; connected objects whose area is greater than a first threshold and whose angle between the long axis of the connected object and the offshore direction of the target location is less than a second threshold are retained and determined as possible rift units;

[0097] The possible rift unit tracking unit is used to track possible rift units at multiple moments. For each possible rift unit at each moment, the remaining possible rift units are its candidate units, and a candidate unit closest to the possible rift unit in the offshore direction is selected, and both are marked as rifts with the same serial number; the possible rift unit tracking unit is repeatedly executed until all possible rift units have been marked with serial numbers;

[0098] The rip flow identification and verification unit is used to superimpose each rip flow number on the flow field diagram to draw its minimum external bounding box, and obtain the rip flow identification result through manual verification.

[0099] The computer-readable storage media may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, flash memory, or any other medium that can be used to store program code in the form of instructions or data structures and that can be accessed by a computer.

[0100] The processor is used to execute the computer program stored in the memory to implement each step of the method involved in the above embodiment.

Claims

1. A method for automatic identification of rip current based on flow field data, characterized in that: It includes the following steps: (1) Establish a grid along the offshore direction and the alongshore direction at the target location, and obtain the flow velocity and flow direction of each grid point at each moment. (2) Search for possible rip current cells at each moment. For grid points where all flow velocities and flow directions are within the threshold range, extract connected objects and calculate the area and minimum bounding box of each connected object; retain connected objects with an area greater than the first threshold and an angle between the major axis of the connected object and the offshore direction of the target location less than the second threshold, and determine them as possible rip current cells. (3) Track possible rip current cells at multiple moments. For each possible rip current cell at each moment, the other possible rip current cells are its candidate cells. Select the candidate cell that is the closest to the possible rip current cell in the offshore direction, and mark the two as rip currents with the same serial number; repeat step (3) until all possible rip current cells have been marked with serial numbers. (4) Superimpose and draw the minimum bounding box of each serial number of rip current on the flow field diagram, and obtain the rip current identification result through manual verification.

2. The method for automatic rip flow identification based on flow field data according to claim 1, characterized in that: In step (1), the flow field data of the target location is obtained and the time-averaged velocity field is calculated. A grid is established along the offshore direction and the coastal direction. The offshore velocity u and the coastal velocity v of each grid point are obtained by linear interpolation. The velocity of each grid point is calculated. and flow direction 3. The method for automatic rip flow identification based on flow field data according to claim 1, characterized in that: In step (2), extracting connected objects and calculating the area and minimum bounding box of each connected object includes: If the flow velocity s and flow direction θ of the grid point satisfy s≥U and θ<α, then generate a mask with a mask value of 1 for this grid point, otherwise generate a mask with a mask value of 0; U is the rip current velocity threshold, and α is the rip current flow direction threshold. Use 8-connected neighborhood connectivity to extract connected objects in the mask, calculate the area of each connected object, and the minimum bounding box of the connected object.

4. The method for automatic rip flow identification based on flow field data according to claim 1, characterized in that: In step (2), before retaining connected objects with an area greater than the first threshold and an angle between the major axis of the connected object and the offshore direction of the target location less than the second threshold, it also includes: Determine whether the area a of the connected object satisfies a ≥ B 2 , if it is not satisfied, the connected object will be deleted; B is the geometric threshold; Calculate the minimum distance d between the minimum bounding boxes of two connected objects, merge two connected objects that satisfy d<B, and update the area and minimum bounding box of the connected object. Delete the connected object that is nearly parallel to the shoreline on the side close to the shoreline of the connected object, and update the area and minimum bounding box of the connected object.

5. The method for automatic identification of rip current based on flow field data according to claim 4, characterized in that: The determination of whether the connected object contains a connected object with a boundary parallel to the shoreline, and if so, deleting the connected object and updating the area and minimum bounding box of the connected object includes: (5.1) For the connected object S, bisect its minimum bounding box along the alongshore direction to obtain two connected objects S1 and S2, where S1 is the connected object close to the shoreline. (5.2) Calculate the angles between the major axes of S1 and S2 and the offshore direction respectively. If the angle is greater than α, then take the length of the major axis of the connected object as its alongshore characteristic width, otherwise take the length of its minor axis, and obtain the characteristic widths w1 and w2 of S1 and S2; α is the rip current flow direction threshold. (5.3) Calculate the median w3 of the width of S2 along the alongshore direction. (5.4) If the determination condition max(w1,w2)>2B and w1>1.5*w3 is satisfied, then move the boundary of the minimum bounding box of S close to the coast one grid point unit in the offshore direction to obtain a new minimum bounding box, and repeat steps (5.1) to (5.3); until the determination condition is not satisfied, or when w1 or w2 is 0, obtain the updated minimum bounding box. (5.5) Use the updated minimum bounding box to extract the updated connected object S'.

6. The method for automatic rip flow identification based on flow field data according to claim 1, characterized in that: In step (2), the connected objects whose reserved area is greater than the first threshold and whose angle between the long axis of the connected object and the offshore direction of the target location is less than the second threshold include: Delete the area a of the connected object that does not satisfy a ≥ B 2 connected objects, B is the geometric threshold; Delete the connected objects whose angle γ between the major axis of the minimum bounding box of the connected object and the offshore direction does not satisfy γ<1.33α, where α is the rip current direction threshold.

7. The method for automatic rip flow identification based on flow field data according to claim 1, characterized in that: In step (3), before selecting a candidate unit that is closest to the possible rip current unit in the offshore direction and marking the two as rip currents with the same serial number, the method further includes: For each possible rift unit at each moment, it is moved by B sinβ / sin45° in the coastal direction and extended by B cosβ / cos45° in the offshore direction. If the overlap area between the minimum bounding box of the candidate rift unit and the moved rift unit is less than 60%, the candidate unit is deleted; β is the average flow direction on the offshore boundary of the possible rift unit, and B is the geometric threshold.

8. The method for automatic rip flow identification based on flow field data according to claim 1, characterized in that: Step (3), selecting a candidate unit that is closest to the possible rift unit in the offshore direction, and marking the two rift units with the same serial number include: For each possible rip current unit at each moment, if its displacement to the candidate unit is not positive in the offshore direction, the candidate unit is deleted; Among the remaining candidate units, the distance between the center point of the minimum external frame of the possible split flow unit and the center point of the minimum external frame of the candidate unit is calculated, the candidate unit with the shortest distance is selected, and the possible split flow unit and the candidate unit are marked as split flows with the same serial number.

9. The method for automatic rip flow identification based on flow field data according to claim 8, characterized in that: The duration of each labeled rip flow is calculated, and the rip flows with a duration less than a third threshold are deleted.

10. A rip flow automatic identification system based on flow field data, characterized in that: include: A data acquisition unit is used to establish a grid at a target location along an offshore direction and a coastal direction, and obtain a flow velocity and a flow direction at each grid point at each moment; A possible rift unit search unit is used to search for possible rift units at each moment. For all grid points whose flow velocity and flow direction are within the threshold range, connected objects are extracted and the area and minimum bounding box of each connected object are calculated; connected objects whose area is greater than a first threshold and whose angle between the long axis of the connected object and the offshore direction of the target location is less than a second threshold are retained and determined as possible rift units; The possible rift unit tracking unit is used to track possible rift units at multiple moments. For each possible rift unit at each moment, the remaining possible rift units are its candidate units, and a candidate unit closest to the possible rift unit in the offshore direction is selected, and both are marked as rifts with the same serial number; the possible rift unit tracking unit is repeatedly executed until all possible rift units have been marked with serial numbers; The rip flow identification and verification unit is used to superimpose each rip flow number on the flow field diagram to draw its minimum external bounding box, and obtain the rip flow identification result through manual verification.

Images