Urban waterlogging diffusion direction determination method based on terrain flow direction and multi-temporal accumulated water boundary cooperation
By combining topographic flow direction and multi-temporal water accumulation boundaries, the logical errors and anti-interference problems in determining the direction of urban flooding diffusion are solved, providing a fast and accurate determination of diffusion direction, which is suitable for emergency response.
Patent Information
- Application Number
- CN202611550002.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-08
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies cannot simultaneously meet the requirements of determining the direction of urban flooding spread in emergency scenarios, including correct physical logic, strong anti-interference ability, minute-level output, and good interpretability. In particular, the calculation is unstable under the constraints of small terrain slope and curbs in urban roads.
Based on the method of combining topographic flow direction and multi-temporal water accumulation boundary, the flow direction is calculated by topographic data, and candidate diffusion directions are generated by combining curb constraints and multi-temporal water accumulation boundary verification, and pseudo-displacement interference is eliminated to determine the dominant diffusion direction.
It enables rapid, accurate, and interpretable determination of the direction of urban flooding spread, avoiding causal inversion and spurious displacement noise interference, and is suitable for emergency response needs.
Smart Images

Figure CN123119293A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of urban flooding monitoring and risk assessment technology, specifically involving a method for determining the direction of urban flooding diffusion based on high-precision topographic flow direction analysis and multi-temporal water accumulation boundary coordination. It is applicable to the rapid prediction and emergency command of the subsequent diffusion trend of urban road water accumulation. Background Technology
[0002] Urban flooding caused by heavy rainfall has become a frequent disaster threatening urban public safety. Quickly and accurately determining the subsequent spread direction and high-risk paths of floodwater is crucial for developing traffic control measures, evacuating people, and pre-deploying rescue forces, directly determining the efficiency of emergency response and the extent of disaster damage.
[0003] Current technologies for determining the direction of urban flooding spread mainly fall into two categories: The first type is the manual experience-based judgment method, which involves on-site inspectors making subjective inferences based on road slope, water accumulation patterns, and historical disaster situations. This method requires no complex equipment, but it has three major fatal flaws: First, it is highly subjective and lacks consistency, with different people's judgments on the same scenario not being consistent enough; Secondly, the coverage area is limited. Due to factors such as low visibility and deep water accumulation caused by heavy rain, it is impossible for personnel to enter dangerous areas to conduct full-area patrols. Third, the timeliness is insufficient. Generally, a single person can only cover less than 10km of roads per day, which cannot meet the needs of simultaneous assessment of large-scale urban flooding.
[0004] The second category is hydrodynamic model simulation, which uses basic data such as topographic elevation, drainage network, rainfall intensity, and road surface roughness to construct a fluid dynamics model and simulate the global diffusion process of accumulated water. Theoretically, this method can output detailed spatiotemporal distribution results, but its practical engineering applicability is extremely poor. First, it has an extremely high data dependence, requiring more than ten types of data such as high-precision road surface DEM, real-time pipeline flow, and manhole distribution. Many cities across the country lack this kind of refined basic data. Second, the calculation time is too long. Even with complete data, full-domain road network simulation generally takes several hours, which cannot meet the command and decision-making requirements of "results in 5 minutes" in emergency scenarios. Third, model parameter calibration is difficult. Parameters such as the roughness of road surface water and the flow capacity of drainage outlets are greatly affected by on-site conditions, and simulation results without actual measurement data calibration have a high error rate.
[0005] In recent years, the development of UAV remote sensing and image segmentation technologies has enabled minute-level, full-area acquisition of urban road flood boundaries, providing a new data foundation for determining the direction of flood spread. However, existing image-based judgment techniques suffer from a fundamental flaw of inverted physical logic: they generally equate the displacement direction of the flood boundary at different times directly with the future spread direction, completely ignoring the fact that flood spread is essentially a gravity-driven water flow, the direction of which is determined by terrain slope and spatial connectivity, rather than the instantaneous morphological changes of the boundary. In actual emergency scenarios, noise factors such as ripples generated by passing vehicles, surges formed by wind blowing on the water surface, and local suction from drainage outlets can cause a large number of local pseudo-displacements at the flood boundary, resulting in a high misjudgment rate based on such pseudo-displacements. In addition, when the flood is in the slow infiltration or overall drainage stage, the boundary may shrink as a whole, but local areas may still spread to low-lying areas; simply relying on boundary displacement cannot identify such hidden risks.
[0006] Urban roads, as a primary site for urban flooding, possess unique structural characteristics that distinguish them from natural watersheds: First, the road surface has a very small slope (standard cross slope 2%~4%, longitudinal slope 1%~3%), and the elevation noise of conventional precision DEM will cause the flow direction calculation to be completely distorted. Secondly, as a continuous rigid boundary, the curb plays a decisive role in blocking or guiding the spread of accumulated water, with most urban road water being confined to the lane area by the curb. However, existing technologies have not designed specific processing mechanisms for these core scenario characteristics: they have neither solved the problem of unstable flow direction calculation on flat roads nor incorporated curb constraints into the determination logic of the diffusion path, resulting in a serious discrepancy between the determination results and the actual diffusion path.
[0007] In summary, existing technologies cannot simultaneously meet the core requirements of "physically correct logic, strong anti-interference ability, minute-level output, and good interpretability" in urban flooding emergency scenarios. There is an urgent need to develop a flood diffusion direction determination method that uses terrain flow direction as the core driving force, couples curb space constraints, and can be combined with multi-temporal boundary dynamic verification to solve the pain points of existing technologies such as reversed causality, poor anti-interference, and insufficient timeliness. Summary of the Invention
[0008] The purpose of this invention is to provide a method for determining the diffusion direction of urban flooding based on the coordination of topographic flow direction and multi-temporal water accumulation boundaries, so as to solve the problems of existing methods such as inverted physical causality, poor anti-interference ability, and difficulty in balancing speed and accuracy. It can quickly output accurate, reliable and interpretable diffusion direction without complex hydrodynamic modeling.
[0009] This invention employs the following technical solution: a method for determining the direction of urban flooding diffusion based on the synergy of topographic flow direction and multi-temporal water accumulation boundaries, comprising the following steps: Step 1: Data Acquisition and Preprocessing. Acquire high-precision topographic data of the target road area. The topographic data should be a digital elevation model (DEM). Simultaneously, acquire the water accumulation boundary vector data of the same target road area for the first and second time phases.
[0010] Step 2: Generation of candidate diffusion directions driven by terrain flow.
[0011] 2.1 Extracting surface water flow direction from the DEM for the target road area. When the local grid slope is greater than 1°, the water flow tends to concentrate significantly, and the surface water flow direction is calculated using the D8 single-direction algorithm; when the local grid slope is less than or equal to 1°, the water flow is dispersed, and the surface water flow direction is calculated using the D∞ multi-direction algorithm. The calculation is limited to the road surface area (by removing buildings, green belts, etc., using road network vectors). Considering that urban road slopes are typically small (cross slope 2%–4%, longitudinal slope 1%–3%), DEM elevation noise may cause instability in flow direction calculations. When the average slope of the target area is less than a preset slope threshold (e.g., 2°), pre-stored road design cross and longitudinal slope data are introduced to correct the DEM flow direction, or the design slope is directly used to replace the DEM flow direction calculation, ensuring the physical rationality of the flow direction results.
[0012] 2.2 Extracting Outflow Edge Points. Unlike existing methods that use boundary curvature for screening, this invention determines the outflow edge point based on the relationship between flow direction and boundary geometry: For each point on the water accumulation boundary contour, the flow direction vector of the corresponding DEM grid is determined, and the outward normal vector (pointing outward from the water body) of that point on the water accumulation boundary is calculated. The angle between the flow direction vector and the outward normal vector is calculated. When this angle is less than a preset angle threshold (usually set to 30° to 45°) and the grid slope is greater than a preset slope threshold (usually 1° to 3°), it indicates that the water flow tends to flow outward from that point, and this point is marked as an outflow edge point.
[0013] 2.3 For each outflow edge point, a local diffusion direction vector is generated along its flow direction. Angle clustering is performed on all local diffusion direction vectors. Due to the constraints of curbs and buildings on urban road structures, water outflow points mainly correspond to road extension directions (such as straight ahead, crossroads, or T-junction branches). The number of clusters for angle clustering is set to 1 to 3, and the direction of the cluster center is the candidate diffusion direction, forming a candidate diffusion direction set. The direction density weight of each candidate direction is defined as the ratio of the number of outflow edge points belonging to that cluster to the total number of outflow edge points. The physical meaning of this weight is: what proportion of all water outflow points points point in that direction, reflecting the "support" of the terrain for that direction.
[0014] Step 3: Connectivity filtering based on curb constraints.
[0015] 3.1 Obtain curbstone vector data, including continuous curbstone line features and opening point features (such as intersections, pedestrian crossing gaps, drainage outlets, etc.).
[0016] 3.2 For each direction in the candidate diffusion direction set, starting from the outflow edge point corresponding to that direction (rather than the centroid of the water accumulation boundary), draw a buffer ray with a certain width (e.g., 2m) along that direction, extending to a preset maximum distance (e.g., 100m). Determine whether this buffer intersects with the continuous curb line element: if they intersect and no opening point is located within the intersection segment, then the flow path in that direction is completely blocked by the curb, the candidate direction is invalid, and it is removed from the candidate set.
[0017] 3.3 Retain all directions that pass the connectivity test to form a set of candidate directions for terrain connectivity.
[0018] Step 4: Dynamic verification of multi-temporal boundaries.
[0019] 4.1 The time interval between the two phases is determined based on the water diffusion rate, which is usually 3 to 15 minutes. When the longitudinal slope of the road is greater than 2%, the water flow is faster and the smaller value is taken; when it is less than 1%, the larger value is taken.
[0020] 4.2 For candidate directions of terrain connectivity, a fan-shaped region is defined with the outflow edge point corresponding to that direction as the vertex, the center line as the center line, and the included angle as the preset verification angle (e.g., ±15°).
[0021] 4.3 Calculate the relative rate of change of the water accumulation area in the second time phase relative to the first time phase within the sector region: Rate of change = (Water accumulation area in the sector region in the second time phase - Water accumulation area in the sector region in the first time phase) / (Water accumulation area in the sector region in the first time phase + Preset minimum positive number). The preset minimum positive number is used to prevent division by zero errors when there is no water accumulation at the initial time.
[0022] 4.4 If the rate of change is less than or equal to zero, it indicates that the water accumulation in that direction has not expanded or is even contracting, and the candidate direction is then eliminated from the candidate directions for terrain connectivity; if the rate of change is greater than zero, it is retained. This design uses temporal change as a rejection condition rather than a weighting term, avoiding the problems of dimensional mismatch and absurd physical meaning of negative values caused by direct weighted summation.
[0023] Step 5: Determine the direction of advantage diffusion.
[0024] 5.1 If, after the third or fourth step, only one direction remains in the candidate direction set for terrain connectivity, then that direction is directly identified as the dominant diffusion direction.
[0025] 5.2 If multiple directions remain, the direction with the highest directional density weight is selected as the dominant diffusion direction. The directional density weight itself already incorporates the statistical significance of the number of terrain outflow points, and its physical meaning is clear, so there is no need to introduce a composite weight that is difficult to calibrate.
[0026] Step 6: Result verification and output.
[0027] 6.1 Verify the rationality of the dominant diffusion direction: Verify whether it is consistent with the overall longitudinal slope of the road (angle less than 45°) and whether there is a continuous linear connected path in the road structure (not blocked by buildings or other impassable obstacles). If the verification fails, return to step two, appropriately reduce the angle threshold or slope threshold, and regenerate the candidate diffusion direction set.
[0028] 6.2 After successful verification, the main diffusion direction is output. Along the main diffusion direction, continuous lane sections that show a decreasing trend relative to the average road elevation are extracted using the DEM and road network topology. These are used as high-risk diffusion channels and output as vector lines and surface features, which are directly overlaid on the electronic map for emergency command use.
[0029] Furthermore, the present invention also provides an apparatus, electronic device, and storage medium for implementing the above method, which specifically comprises the following interconnected modules: (1) Data acquisition module: used to acquire high-precision terrain data, curbstone vector data, and water boundary vector data of the first and second time phases of the same target road area; (2) Flow direction analysis module: It is connected to the data acquisition module to receive terrain data and water boundary data, calculate local flow direction based on terrain data, determine outflow edge point by combining the outflow normal of water boundary, and perform angle clustering on local diffusion direction to generate candidate diffusion direction set, while calculating density weight of each direction. (3) Constraint filtering module: It is connected to the flow direction analysis module and the data acquisition module to receive the candidate diffusion direction set and curb vector data. Starting from the outflow edge point, it performs spatial connectivity filtering based on the curb boundary, eliminates the blocked directions, and obtains the terrain connectivity candidate direction set. (4) Dynamic verification module: It is connected to the constraint screening module and is used to receive the set of candidate directions for terrain connectivity, calculate the relative change rate of water area in the preset sector area for each candidate direction using the multi-temporal water accumulation boundary, and eliminate directions whose area has not expanded. (5) Direction decision module: It is communicatively connected to the dynamic verification module and is used to determine the dominant diffusion direction from the verified candidate directions according to the direction density weight; (6) Result output module: It is connected to the direction decision module and is used to verify the rationality of the dominant diffusion direction. After the verification is passed, the main diffusion direction is output, and the high-risk diffusion channel is output in combination with the road topology.
[0030] The electronic device includes a memory, a processor, and a communication interface. The memory stores a computer program that implements the aforementioned determination method; the processor is communicatively connected to the memory and is used to call and execute the computer program to achieve complete logical control from data acquisition to result output; the communication interface is used to receive data transmitted from the UAV and issue early warning commands. This electronic device can be specifically implemented as a ground-based emergency command workstation, a vehicle-mounted processing terminal, or a cloud-based command server.
[0031] The computer-readable storage medium stores a computer program that, when executed by a processor of an electronic device, enables the implementation of the steps involved in determining the direction of urban flooding spread. The storage medium may include a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), or cloud storage.
[0032] Compared with the prior art, the present invention has the following significant advantages: 1. Correct physical logic: The terrain slope is used as the active driving force for diffusion to generate candidate directions. The outflow edge point is used as the starting point for spatial analysis, the curb is used as a constraint, and the time series image is used as the rejection verification condition. This corrects the previous error of "inferring the trend from the image change" and the judgment result is consistent with the basic law of water flowing downhill.
[0033] 2. Strong anti-interference capability: It does not rely on boundary point displacements that are difficult to calculate accurately, thus avoiding pseudo-displacement noise caused by wind, waves, and vehicle disturbances. Even with only a single-phase water accumulation boundary, it can provide a reasonable prediction based on the terrain; multi-phase data is used as a veto condition, avoiding dimensional conflicts caused by direct linear weighting.
[0034] 3. Fast and low cost: No need to build a hydrodynamic model, the calculation time is only a few seconds to a minute, and the data requirements are only DEM and water accumulation outline, which is very suitable for emergency sites.
[0035] 4. High interpretability: The output of the main diffusion direction and high-risk channels has clear topographical, connectivity, and area change data, which emergency command personnel can directly use for traffic control and personnel evacuation without requiring professional background to understand and adopt. For example, when the device of this invention is deployed on an emergency command server, the data acquisition module receives real-time data from the UAV on the water accumulation boundary and pre-stored DEM; the flow direction analysis module completes outflow point extraction and direction clustering in only 0.5 seconds; the constraint screening module completes curbstone query in only 0.1 seconds; the dynamic verification module completes area change rate calculation in only 0.3 seconds; and the direction decision module outputs the results instantly. The entire process is automated and can process multiple roads simultaneously, greatly improving the speed of emergency response. Attached Figure Description
[0036] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a flowchart illustrating the overall process of the method of the present invention. Figure 2 A schematic diagram for extracting outflow edge points (determining the angle between the flow direction and the outward normal direction); Figure 3 This is a schematic diagram of curbstone connectivity screening starting from the outflow edge point; Figure 4 A schematic diagram of the sector region for multi-temporal dynamic verification; Figure 5 This is a schematic diagram of the module structure of the device of the present invention. Detailed Implementation
[0037] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.
[0038] The following, in conjunction with the accompanying drawings, details the method for determining the urban flooding diffusion direction based on the coordination of topographic flow direction and multi-temporal water accumulation boundaries provided by various embodiments of the present invention, including the following steps: The first step is data acquisition: acquiring high-precision terrain data of the target road area, as well as water boundary vector data of the first and second time phases of the same target road area; The second step is the generation of candidate directions driven by terrain flow: Based on the terrain data, the surface water flow direction of the target road area is extracted; when the average slope of the area is less than a preset slope threshold, the flow direction is calculated by correcting or replacing the DEM calculation with pre-stored road design cross slope and longitudinal slope data; for each point on the water accumulation boundary contour, its corresponding surface water flow direction vector is determined, and the angle between the flow direction vector and the outward normal vector of the water accumulation boundary is calculated. Points with an angle less than a preset angle threshold and a grid slope greater than a preset slope threshold are selected as outflow edge points; for each outflow edge point, its local diffusion direction vector is obtained, and all local diffusion direction vectors are clustered by angle to form 1 to 3 candidate diffusion directions, constituting a candidate diffusion direction set; the direction density weight G of each candidate diffusion direction d is defined. d n is the number of outflow edge points in the cluster corresponding to this candidate direction. d The ratio of the total number of outflow edge points N to the total number of outflow edge points, i.e. ; The third step is to filter the connectivity of curb constraints: obtain the curb boundary vector data of the target road area, including continuous curb segments and opening positions; for each candidate direction in the candidate diffusion direction set, take the outflow edge point corresponding to the candidate direction as the starting point, and determine whether the path is completely blocked by continuous curbs and has no openings along the direction; if so, remove the candidate direction from the candidate set; retain the remaining candidate directions as the set of terrain connectivity candidate directions. Step 4, Multi-temporal Dynamic Verification: For each candidate direction in the set of candidate directions for terrain connectivity, define a sector region with the outflow edge point corresponding to the candidate direction as the vertex, the candidate direction as the center line, and the included angle as the preset verification angle; calculate the relative change rate of the water accumulation area in the second temporal phase relative to the first temporal phase within the sector region; if the relative change rate of the water accumulation area is less than or equal to zero, it indicates that the water accumulation has not expanded or is contracting in that direction, and then the candidate direction is removed from the set of candidate directions for terrain connectivity. Step 5: Determining the dominant diffusion direction: If only one candidate direction exists after steps 3 or 4, it is directly determined as the dominant diffusion direction; if multiple directions exist, the direction density weight G is selected. d The largest candidate direction is taken as the direction of advantage diffusion. Step 6, Output Results: After the dominant diffusion direction has been validated for rationality, it is output as the main diffusion direction. Along the main diffusion direction, combined with the road connectivity structure, sections that show a decreasing trend relative to the average road elevation and are linearly connected are output as high-risk diffusion channels. The rationality validation includes: determining whether the main diffusion direction is consistent with the overall longitudinal slope of the road and whether there is a continuous linearly connected path in the road structure. If either condition is not met, a parameter backtracking loop is entered: the preset angle threshold and preset slope threshold from step 2 are decreased by a preset step size, and the process returns to step 2 to regenerate the candidate diffusion direction set. If the validation still fails after a preset maximum number of loops or after the threshold is reduced to a preset lower limit, the candidate direction with the highest current direction density weight is directly output and a low-confidence label is added.
[0039] In the first step, the high-precision terrain data includes at least ground elevation information, with a planar resolution better than 1m and an elevation accuracy better than 0.1m. The high-precision terrain data comes from vehicle-mounted lidar point clouds, UAV oblique photogrammetry, or existing urban high-precision DEMs, and its planar coordinate system is consistent with the water accumulation boundary vector data. The water accumulation boundary vector data is obtained by UAV visible light image segmentation or by dual-modal fusion segmentation of visible light and thermal infrared images.
[0040] In the second step, when the local grid slope is greater than 1°, the water flow tends to concentrate significantly, and the surface water flow direction is calculated using the D8 single-direction algorithm; when the local grid slope is less than or equal to 1°, the water flow is dispersed, and the surface water flow direction is calculated using the D∞ multi-direction algorithm; the accessible area of the road is limited to the road surface area, excluding inaccessible areas such as buildings and medians.
[0041] In the second step, angle clustering is performed on all local diffusion direction vectors to form candidate diffusion directions. Since the urban road structure is restricted by curbs and buildings, the water outflow outlets mainly correspond to the road extension direction. The number of clusters for angle clustering is set to 1 to 3. The preset angle threshold is 30° to 45°, and the preset slope threshold is 1° to 3°. The preset angle threshold and preset slope threshold adopt adaptive adjustment logic: the angle threshold is dynamically adjusted according to the elevation roughness of the DEM of the target road area. The larger the elevation roughness, the larger the angle threshold. The slope threshold is dynamically adjusted according to the average longitudinal slope of the target road area. The larger the average longitudinal slope, the larger the slope threshold.
[0042] In the third step, the method for determining whether the path is blocked by continuous curbs is as follows: starting from the outflow edge point, draw a buffer ray of a preset width along the candidate direction and extend it to a preset maximum distance; if the buffer ray intersects with the line element representing the continuous curb within this distance, and there is no curb opening element in the intersection segment, then it is determined to be blocked.
[0043] In the first step, the water accumulation boundaries of different time phases are geographically registered based on the road centerline and fixed landmarks at intersections; in the fourth step, the time interval between the first and second time phases is determined according to the water diffusion rate, with a value ranging from 3 to 15 minutes; when the road longitudinal slope is greater than 2%, the smaller value is taken, and when it is less than 1%, the larger value is taken.
[0044] In the fourth step, the formula for calculating the relative change rate of the water accumulation area is: Change rate = (Water accumulation area of the sector region in the second time phase - Water accumulation area of the sector region in the first time phase) / (Water accumulation area of the sector region in the first time phase + Preset minimum positive number) Example 1: Determining based solely on the current water accumulation boundary and terrain A main road in a city experienced flooding, with a standard curb height of 18cm. A visible light boundary image of the flooded area and a DEM (Diagram of the flooded area) with an accuracy of 0.5m, along with curb vector data, have been acquired from a UAV at the current time. The DEM shows that the road's longitudinal slope slopes downstream by approximately 2.5%, and its cross slope slopes to the right by approximately 2%.
[0045] The second step involves using the D8 algorithm to calculate the flow direction in the road area, and introducing a road design slope correction for flat sections with an average slope of less than 2°. The angle between the flow direction and the outward normal at each point on the water accumulation boundary is calculated, with an angle less than 40° and a slope greater than 1.5° as thresholds, resulting in eight outflow edge points. After clustering, two candidate directions are formed: direction 1 (downstream of the road, n1=6) and direction 2 (lateral towards the sidewalk opening, n2=2). The direction density weights are G1=6 / 8=0.75 and G2=2 / 8=0.25, respectively.
[0046] The third step involves drawing buffer rays starting from the outflow edge points corresponding to directions 1 and 2, respectively. The ray from direction 2 intersects with a continuous curb at 15m, and since this section has no openings, it is considered blocked and discarded. Direction 1, however, has intersections along its route and is not blocked, so it is retained. At this point, only direction 1 remains as a candidate direction for terrain connectivity. Since only one candidate direction for climate connectivity remains, and there is no multi-temporal data, direction 1 is directly determined as the dominant diffusion direction. Verification confirms that direction 1 aligns with the road's longitudinal slope and has a continuous connectivity path, thus passing the verification. Direction 1 is output as the primary diffusion direction, and the high-risk channel is the right lane of the downstream section, where the elevation decreases significantly relative to the road's average elevation. Subsequent on-site observations confirm that the water accumulation does indeed spread downstream, confirming the correct determination.
[0047] Example 2: Determination using dual-temporal boundary enhancement In the above scenario, the water accumulation boundary was also obtained 5 minutes ago and at present (because the longitudinal slope is greater than 2%, the time interval is taken as a smaller value of 3 to 5 minutes).
[0048] The fourth step is to calculate the relative rate of change of area for the fan-shaped region in direction 1 (with the corresponding outflow edge point as the vertex, the centerline as direction 1, and the included angle ±15°). According to GIS statistics, the water accumulation area in the fan-shaped region during the first time phase was 120 square meters, and the water accumulation area in the fan-shaped region during the second time phase was 180 square meters. Substituting into the calculation formula: Rate of change = (180-120) / (120+0.001) ≈ 0.50. This indicates that the water accumulation in this direction is significant, and direction 1 is retained after dynamic verification. Direction 2 was eliminated in the third step and does not need to be calculated.
[0049] As a comparative experiment, the same test was conducted on another adjacent uphill section of road. It was found that the area in the first phase of a candidate direction on this section was 80 square meters, and the area in the second phase was 75 square meters (due to localized leakage). The rate of change was (75-80) / (80+0.001) ≈ -0.06. Because the rate of change was less than zero, this direction was successfully eliminated, avoiding the error of misjudging the direction of water receding as the direction of diffusion.
[0050] Fifth, direction 1 is the only retained direction and is determined as the dominant diffusion direction. The main diffusion direction is output, and the high-risk channel is the right lane of the downstream road segment with a significant downward trend relative to the average road elevation. In this example, dynamic verification strengthens the confidence in direction 1.
[0051] The above description is merely an embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.
Claims
1. A method for determining the direction of urban flooding diffusion based on the synergy of topographic flow direction and multi-temporal water accumulation boundaries, characterized in that, Includes the following steps: The first step is data acquisition: acquiring high-precision terrain data of the target road area, as well as water boundary vector data of the first and second time phases of the same target road area; The second step is the generation of candidate directions driven by terrain flow: Based on the terrain data, the surface water flow direction of the target road area is extracted; when the average slope of the area is less than a preset slope threshold, the flow direction is calculated by correcting or replacing the DEM calculation with pre-stored road design cross slope and longitudinal slope data; for each point on the water accumulation boundary contour, its corresponding surface water flow direction vector is determined, and the angle between the flow direction vector and the outward normal vector of the water accumulation boundary is calculated. Points with an angle less than a preset angle threshold and a grid slope greater than a preset slope threshold are selected as outflow edge points; for each outflow edge point, its local diffusion direction vector is obtained, and all local diffusion direction vectors are clustered by angle to form 1 to 3 candidate diffusion directions, constituting a candidate diffusion direction set; the direction density weight G of each candidate diffusion direction d is defined. d n is the number of outflow edge points in the cluster corresponding to this candidate direction. d The ratio of the total number of outflow edge points N to the total number of outflow edge points, i.e. ; The third step is to filter the connectivity of the curb constraint: obtain the curb boundary vector data of the target road area, including continuous curb segments and opening locations; For each candidate direction in the candidate diffusion direction set, starting from the outflow edge point corresponding to the candidate direction, determine whether the path is completely blocked by continuous curb stones and has no opening along the direction. If so, then remove that candidate direction from the candidate set; The remaining candidate directions are retained as a set of candidate directions for terrain connectivity; Step 4, Multi-temporal Dynamic Verification: For each candidate direction in the set of candidate directions for terrain connectivity, define a sector region with the outflow edge point corresponding to the candidate direction as the vertex, the candidate direction as the center line, and the included angle as the preset verification angle; calculate the relative change rate of the water accumulation area in the second temporal phase relative to the first temporal phase within the sector region; if the relative change rate of the water accumulation area is less than or equal to zero, it indicates that the water accumulation has not expanded or is contracting in that direction, and then the candidate direction is removed from the set of candidate directions for terrain connectivity. Step 5: Determining the dominant diffusion direction: If only one candidate direction exists after steps 3 or 4, it is directly determined as the dominant diffusion direction; if multiple directions exist, the direction density weight G is selected. d The largest candidate direction is taken as the direction of advantage diffusion; Step 6, Output Results: After the rationality verification of the dominant diffusion direction, output the main diffusion direction. Then, along the main diffusion direction and combined with the road connectivity structure, output the sections that show a decreasing trend relative to the average road elevation and are linearly connected as high-risk diffusion channels. The rationality verification includes: determining whether the main diffusion direction is consistent with the overall direction of the road longitudinal slope and whether there is a continuous linear connected path in the road structure; If any condition is not met, the parameter backtracking loop is entered: the preset angle threshold and preset slope threshold in the second step are reduced by a preset step size, and the second step is returned to regenerate the candidate diffusion direction set; if the verification still fails after the preset maximum number of loops or the threshold is reduced to the preset lower limit, the candidate direction with the largest current direction density weight is directly output and a low confidence label is attached.
2. The method for determining the urban flooding diffusion direction based on the synergy of topographic flow direction and multi-temporal water accumulation boundaries as described in claim 1, characterized in that, In the first step, the high-precision terrain data includes at least ground elevation information, with a planar resolution better than 1m and an elevation accuracy better than 0.1m. The high-precision terrain data comes from vehicle-mounted lidar point clouds, UAV oblique photogrammetry, or existing urban high-precision DEMs, and its planar coordinate system is consistent with the water accumulation boundary vector data. The water accumulation boundary vector data is obtained by UAV visible light image segmentation or by dual-modal fusion segmentation of visible light and thermal infrared images.
3. The method for determining the urban flooding diffusion direction based on the synergy of topographic flow direction and multi-temporal water accumulation boundaries as described in claim 1, characterized in that, In the second step, when the local grid slope is greater than 1°, the water flow tends to concentrate significantly, and the surface water flow direction is calculated using the D8 single-direction algorithm; when the local grid slope is less than or equal to 1°, the water flow is dispersed, and the surface water flow direction is calculated using the D∞ multi-direction algorithm; the accessible area of the road is limited to the road surface area, excluding inaccessible areas such as buildings and medians.
4. The method for determining the urban flooding diffusion direction based on the synergy of topographic flow direction and multi-temporal water accumulation boundaries as described in claim 1, characterized in that, In the second step, angle clustering is performed on all local diffusion direction vectors to form candidate diffusion directions. Since the urban road structure is restricted by curbs and buildings, the water outflow outlets mainly correspond to the road extension direction. The number of clusters for angle clustering is set to 1 to 3. The preset angle threshold is 30° to 45°, and the preset slope threshold is 1° to 3°. The preset angle threshold and preset slope threshold adopt adaptive adjustment logic: the angle threshold is dynamically adjusted according to the elevation roughness of the DEM of the target road area. The larger the elevation roughness, the larger the angle threshold. The slope threshold is dynamically adjusted according to the average longitudinal slope of the target road area. The larger the average longitudinal slope, the larger the slope threshold.
5. The method for determining the urban flooding diffusion direction based on the synergy of topographic flow direction and multi-temporal water accumulation boundaries as described in claim 1, characterized in that, In the third step, the method to determine whether the path is blocked by continuous curbs is: starting from the outflow edge point, draw a buffer ray of preset width along the candidate direction and extend it to the preset maximum distance; If the buffer zone ray intersects with the line element representing a continuous curb within this distance, and there is no curb opening element within the intersection segment, it is determined to be an obstruction.
6. The method for determining the urban flooding diffusion direction based on the synergy of topographic flow direction and multi-temporal water accumulation boundaries as described in claim 1, characterized in that, In the first step, the water accumulation boundaries of different time phases are geographically registered based on the road centerline and fixed landmarks at intersections; in the fourth step, the time interval between the first and second time phases is determined according to the water diffusion rate, with a value ranging from 3 to 15 minutes; when the road longitudinal slope is greater than 2%, the smaller value is taken, and when it is less than 1%, the larger value is taken.
7. The method for determining the urban flooding diffusion direction based on the synergy of topographic flow direction and multi-temporal water accumulation boundaries as described in claim 1, characterized in that, In the fourth step, the formula for calculating the relative change rate of the water accumulation area is: Change rate = (Water accumulation area of the fan-shaped region in the second time phase - Water accumulation area of the fan-shaped region in the first time phase) / (Water accumulation area of the fan-shaped region in the first time phase + preset minimum positive number).
8. An urban flooding diffusion direction determination device for implementing the urban flooding diffusion direction determination method according to any one of claims 1 to 7, characterized in that, include: The data acquisition module is used to acquire high-precision terrain data, curbstone vector data, and the water accumulation boundaries of the same target road area in the first and second time phases. The flow direction analysis module is used to receive high-precision terrain data and water accumulation boundary transmitted by the data acquisition module, calculate local flow direction based on terrain, determine outflow edge point by combining boundary outward normal, generate candidate diffusion direction set and calculate density weight of each direction. The constraint filtering module is used to receive the candidate diffusion direction set output by the flow direction analysis module and the curbstone vector data transmitted by the data acquisition module. Starting from the outflow edge point, it performs connectivity filtering based on the curbstone boundary to obtain candidate terrain connectivity directions. The dynamic verification module is used to receive the candidate terrain connectivity directions output by the constraint filtering module, calculate the relative change rate of water accumulation area of the candidate directions using multi-temporal water accumulation boundaries, and eliminate directions whose area has not increased or decreased. The direction decision module is used to determine the direction of advantage diffusion by combining terrain connectivity and dynamic verification results; The result output module is used to receive the dominant diffusion direction output by the direction decision module, verify and output the main diffusion direction and the section that shows a downward trend relative to the average road elevation and is linearly connected as a high-risk diffusion channel.
9. An electronic device, characterized in that, The device includes a memory, a processor, and a computer program stored in the memory. When the processor executes the computer program, it implements the method for determining the direction of urban flooding spread as described in any one of claims 1 to 7. The electronic device is a ground workstation, an emergency command server, or a cloud processing platform.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for determining the direction of urban flooding spread as described in any one of claims 1 to 7.
