Road rescue path navigation method and system

Abstract

The invention relates to the technical field of path planning, and discloses a road rescue path navigation method and system.The road rescue path navigation method comprises the following steps that S101, the optimal time lag is obtained through calculation; step S102, calculating to obtain a line consistency coefficient and a line component average speed; step S103, calculating to obtain a lagging water supply ratio; step S104, calculating to obtain a longitudinal overflow risk index; and S105, performing state marking on the road section according to the longitudinal overflow risk index. According to the method, the water flow direction and speed are analyzed through the light flow, and the influence of the longitudinal slope urgency and blockage state dynamic propagation risk is combined, so that the rescue path planning can effectively avoid potential risk road sections such as accumulated water overflow and can also avoid excessive detour, the timeliness of the rescue path is improved, safer and more efficient passage guidance is provided for rescue actions, and the rescue efficiency is improved. And the traffic safety and the rescue efficiency of rescue workers and equipment in a complex disaster scene are ensured.

Description

Technical Field

[0001] This invention relates to the field of route planning technology, and more specifically, to a road rescue route navigation method and system. Background Technology

[0002] With the increasing frequency of extreme weather events, disasters such as landslides and rockfalls triggered by heavy rainfall are on the rise. These disasters often cause blockages in road culverts and ditches, as well as flooding on roads, posing a severe challenge to the planning of road rescue routes. Rescue operations have extremely high requirements for the timeliness and safety of routes, and must accurately avoid risky sections of road such as those with water accumulation, interruptions, and overflows.

[0003] Existing road rescue route planning mostly relies on static road networks or real-time traffic data, which has obvious limitations. First, there is a natural time lag between rainfall and road surface water accumulation. Rainwater needs a certain process from falling, flowing down the slope, to forming water on the road surface. However, existing methods directly use real-time rainfall data to assess water accumulation risk without calibrating the time difference between the two, which can easily lead to misjudgments such as rainfall decreasing but water accumulation still increasing. Second, culverts and side ditches are key nodes for road drainage, and blockage is the core factor causing water overflow. However, existing technologies lack targeted monitoring of these hotspots, making it impossible to identify blockage status in a timely manner and predict whether water will spread longitudinally along the road.

[0004] Furthermore, existing risk assessments simply determine whether there is water accumulation without considering the characteristics of water flow and light flow in relation to the longitudinal slope of the road. This makes it impossible to quantify the consistency and speed of water flow along the longitudinal direction of the road, and also ignores the impact of the gradient of the longitudinal slope on water accumulation, resulting in the omission of water accumulation risks in downstream sections. These problems often lead to the planned rescue routes being misselected for dangerous sections and excessive detours that prolong rescue time, seriously affecting rescue efficiency and even threatening the safety of rescue personnel and equipment. Summary of the Invention

[0005] This invention provides a road rescue route navigation method and system, which solves the technical problems mentioned in the background.

[0006] This invention provides a road rescue route navigation method, comprising the following steps: Step S101: Collect the rainfall intensity sequence and road image sequence at the accident site, and calculate the water pixel ratio of the road image at each time point to obtain the water pixel ratio sequence. Based on this, perform correlation analysis to obtain the optimal time delay. Step S102: Calculate the optical flow vector of the water pixels in the road image at each time moment, and calculate the consistency coefficient and average velocity of the road components along the route by combining it with the road segment direction vector. Step S103: Calculate the percentage of water pixels in the hotspot area, compare it with a preset threshold to obtain the blockage indication, align the rainfall intensity sequence according to the optimal time lag, and calculate the lag water supply ratio at each moment through a sliding window. Step S104: Based on the consistency coefficient along the line, the average velocity of the components along the line, the blockage indication, and the lag water supply ratio, the longitudinal overflow risk index is calculated by longitudinal slope modulation. Step S105: Mark the status of the road segment according to the longitudinal overflow risk index and mark it on the map at the same time to obtain the road rescue route.

[0007] Furthermore, taking the accident site as the center point and the current location of the rescue team as the boundary point, the line connecting the two points is used as the median line, and a rectangular area is generated with the median line as the fixed width. The rainfall intensity sequence is obtained by extracting the grid radar rainfall information in this area, and the road image sequence is obtained by using a drone to take aerial photos along the line connecting the two points to cover the area.

[0008] Furthermore, the road region in the road image is first identified by the object detection model, then the water region in the road image is identified by the image segmentation model, and finally the number of water pixels falling within the road region is counted and divided by the number of pixels in the road region to obtain the proportion of water pixels in the road image.

[0009] Furthermore, the difference between the element values ​​of adjacent time points in the rainfall intensity sequence and the water body pixel proportion sequence is calculated to obtain the rainfall intensity difference sequence and the water body pixel proportion difference sequence. Then, the inner product sum of the two under different lag coefficients is calculated, and the lag coefficient with the largest inner product sum is taken as the optimal time lag.

[0010] Furthermore, the road image at each time moment is converted into a grayscale image. The Farneback dense optical flow algorithm is used to calculate the optical flow vector of each water pixel at adjacent time moments. If the magnitude of the optical flow vector is less than the noise threshold, the water pixels in the road image at that time moment are removed. Then, the remaining water pixels in the road image at each time moment are statistically averaged to obtain the consistency coefficient and the average velocity of the components along the road.

[0011] Furthermore, the number of water pixels falling within the hotspot area is counted, and the percentage of water pixels in the hotspot area is obtained by dividing the number of pixels in the hotspot area. If this value is greater than or equal to a preset threshold, the blockage indicator is assigned a value of 1; otherwise, it is assigned a value of 0. The hotspot areas are obtained by manually marking them on the road image.

[0012] Furthermore, the original rainfall intensity sequence is aligned by shifting the entire sequence along the time axis by the optimal time lag. A sliding window is applied to each moment of the aligned rainfall intensity sequence. If the element value of the aligned rainfall intensity sequence at each moment within the sliding window is less than the average value of the element values ​​at all moments within the sliding window, the rainfall intensity indicator is assigned a value of 1; otherwise, it is assigned a value of 0. The average value of the rainfall intensity indicator at all moments within the sliding window is then calculated to obtain the lag water supply ratio. The size of the sliding window is a user-defined parameter.

[0013] Furthermore, the minimum value of the weighted fusion result of 1 and the consistency coefficient along the route, the average speed of the component along the route, the congestion indication, and the lag water supply ratio is taken. The four weighted terms involved in the weighted fusion are the product of the first weighted coefficient and the consistency coefficient along the route, the ratio of the average speed of the component along the route to the speed reference multiplied by the second weighted coefficient, 1 divided by 100 divided by the longitudinal slope of the road multiplied by the congestion indication and the third weighted coefficient, and the fourth weighted coefficient and the lag water supply ratio. The first weighted coefficient, the second weighted coefficient, the third weighted coefficient, and the fourth weighted coefficient are all user-defined parameters.

[0014] Furthermore, if the longitudinal overflow risk index is greater than or equal to the first threshold, the road area in the road image is marked as an interrupted road; if the longitudinal overflow risk index is less than the first threshold but greater than or equal to the second threshold, the road area in the road image is marked as a risky road; otherwise, it is marked as a normal road. The first threshold and the second threshold are both user-defined parameters.

[0015] This invention provides a road rescue route navigation system, comprising: The optimal time delay calculation module is used to collect rainfall intensity sequences and road image sequences at the accident site, and to calculate the water pixel ratio sequence of the road images at each time point to obtain the water pixel ratio sequence. Correlation analysis is then performed based on this sequence to obtain the optimal time delay. The parameter calculation module along the route is used to calculate the optical flow vector of the water pixels in the road image at each time moment, and calculate the consistency coefficient and the average velocity of the components along the route by combining it with the road segment direction vector. The lag water supply ratio calculation module is used to calculate the percentage of water pixels in hotspot areas, compare it with a preset threshold to obtain the blockage indicator, align the rainfall intensity sequence according to the optimal time lag, and calculate the lag water supply ratio at each moment through a sliding window. The overflow risk index calculation module is used to calculate the longitudinal overflow risk index by longitudinal slope modulation based on the consistency coefficient along the line, the average velocity of the components along the line, the blockage indication, and the lag supply ratio. The road status marking module is used to mark the status of road segments according to the longitudinal overflow risk index and simultaneously mark them on the map to obtain road rescue routes.

[0016] The beneficial effects of this invention are as follows: By integrating multi-source data such as rainfall, imagery, and road network, and fully considering the time lag between rainfall and water accumulation, the correlation between water flow direction and road longitudinal slope, this invention achieves a quantitative assessment of the longitudinal overflow risk of roads after landslide rainfall. By analyzing water flow direction and velocity through optical flow analysis, and combining the dynamic propagation of risk impacts based on the gradient of longitudinal slope and blockage status, this invention enables rescue route planning to effectively avoid potential risk sections such as water accumulation and flooding, while also avoiding excessive detours. This improves the timeliness of rescue routes, provides safer and more efficient passage guidance for rescue operations, and ensures the safety of rescue personnel and equipment and the efficiency of rescue operations in complex disaster scenarios. Attached Figure Description

[0017] Figure 1 This is a flowchart of a road rescue route navigation method according to the present invention; Figure 2 This is a schematic diagram of a road rescue route navigation system according to the present invention.

[0018] In the diagram: Optimal time delay calculation module 201, along-line parameter calculation module 202, delayed water supply ratio calculation module 203, overflow risk index calculation module 204, and road condition marking module 205. Detailed Implementation

[0019] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0020] It should be noted that, unless otherwise defined, the technical or scientific terms used in one or more embodiments of the present invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in one or more embodiments of the present invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" indicate that the element or object preceding the term encompasses the elements or objects listed following the term and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0021] like Figures 1-2 As shown, a road rescue route navigation method includes the following steps: Step S101: Collect the rainfall intensity sequence and road image sequence at the accident site, and calculate the water pixel ratio of the road image at each time point to obtain the water pixel ratio sequence. Based on this, perform correlation analysis to obtain the optimal time delay. Step S102: Calculate the optical flow vector of the water pixels in the road image at each time moment, and calculate the consistency coefficient and average velocity of the road components along the route by combining it with the road segment direction vector. Step S103: Calculate the percentage of water pixels in the hotspot area, compare it with a preset threshold to obtain the blockage indication, align the rainfall intensity sequence according to the optimal time lag, and calculate the lag water supply ratio at each moment through a sliding window. Step S104: Based on the consistency coefficient along the line, the average velocity of the components along the line, the blockage indication, and the lag water supply ratio, the longitudinal overflow risk index is calculated by longitudinal slope modulation. Step S105: Mark the status of the road segment according to the longitudinal overflow risk index and mark it on the map at the same time to obtain the road rescue route.

[0022] In one embodiment of the present invention, the accident site is taken as the center point and the current location of the rescue team is taken as the boundary point. The line connecting the two points is taken as the middle line. A rectangular range is generated with the middle line as the fixed width. Alternatively, the line connecting the two points can be taken as the diameter to generate a circular range. The rainfall intensity sequence is obtained by extracting the grid radar rainfall information in the area. The road image sequence is obtained by taking aerial photos of the area along the line connecting the two points using a drone.

[0023] In one embodiment of the present invention, the road region in the road image is first identified by an object detection model, then the water region in the road image is identified by an image segmentation model, and finally the number of water pixels falling within the road region is counted and divided by the number of pixels in the road region to obtain the water pixel ratio of the road image. The object detection model can be a YOLO series model or a Faster R-CNN model, etc., and the image segmentation model can be a U-Net model or a Mask R-CNN model, etc.

[0024] It should be noted that the above-mentioned object detection model can be trained using existing public datasets, such as the KITTI dataset and the Cityscapes dataset. Similarly, the above-mentioned image segmentation model can be trained using existing public datasets, such as the FloodNet dataset and the ISAID dataset. In addition, a custom dataset can be constructed, that is, road images covering the application scenario of this invention can be collected, such as damaged roads after landslides and road water accumulation under heavy rainfall. It can also cover different weather conditions and road types. The road and water areas can be labeled using annotation tools such as LabelImg and VGG Image Annotator. Data augmentation can be performed by rotating, flipping, scaling, adjusting brightness and contrast, and adding noise to expand the dataset, thereby improving the generalization ability of the model, especially for complex scenarios such as heavy rain and low light. The dataset is divided into a training set (for model training), a validation set (for parameter adjustment during training), and a test set (for evaluating model performance) in a 7:2:1 ratio, which will not be elaborated here.

[0025] In one embodiment of the present invention, the rainfall intensity sequence and the road image sequence have the same length. For example, rainfall intensity and road images are collected at the same time interval within the same time period. Alternatively, the sequence length can be unified by elimination or interpolation.

[0026] In one embodiment of the present invention, the difference between the element values ​​of adjacent time moments of the rainfall intensity sequence and the water body pixel proportion sequence is calculated to obtain the rainfall intensity difference sequence and the water body pixel proportion difference sequence. Then, the inner product sum of the two under different lag coefficients is calculated, and the lag coefficient with the largest inner product sum is taken as the optimal time lag. The lag coefficient is inner product and The calculation formula is as follows: Where K represents the length of the rainfall intensity difference sequence and the water body pixel proportion difference sequence, Represents the k-th difference sequence of rainfall intensity Each element value This represents the k-th element value of the water body pixel proportion difference sequence. The hysteresis coefficient is a custom auto-incrementing parameter with an initial value of 0. Preferably, the maximum value of the hysteresis coefficient is the floor function of K / 8.

[0027] It should be noted that the water pixel ratio is used to quantify the water accumulation in road areas, which can intuitively reflect the degree of water accumulation at different times and on different road sections. This eliminates the differences caused by cameras with different resolutions or drones with different shooting angles, allowing for direct comparison of water accumulation at different road sections and times. This provides a unified standard of basic data for subsequent analysis of the relationship between rainfall and water accumulation. Since there is a certain delay in the process of rainwater flowing onto the road surface and forming water accumulation after rainfall, such as the time required for rainwater to fall from the air to the ground and flow from the slope to the road, the rainfall intensity sequence and the water pixel ratio sequence are not completely synchronized in time. By determining the best matching time offset between the two through the optimal time lag, the changes in rainfall and changes in water accumulation can be accurately correlated on the timeline. This avoids misjudgments caused by time misalignment, such as rainfall weakening but water accumulation still increasing. This provides a reliable time benchmark for risk assessment of subsequent rescue routes.

[0028] In one embodiment of the present invention, the road image at each time moment is converted into a grayscale image, the Farneback dense optical flow algorithm is used to calculate the optical flow vector of each water pixel at adjacent time moments, and if the magnitude of the optical flow vector is less than the noise threshold, the water pixels in the road image at that time moment are removed. Then, the remaining water pixels in the road image at each time moment are statistically averaged to obtain the consistency coefficient along the line and the average velocity of the components along the line. The consistency coefficient along the road in the remaining water pixels of the road image at time i. The calculation formula is as follows: ; The average velocity of the remaining water pixels along the road in the road image at time i. The calculation formula is as follows: ; in and Let represent the set of remaining water pixels and the number of pixels at time i, respectively. Let d represent the optical flow vector of the j-th remaining water pixel at time i, and let d represent the road segment direction vector. This represents the L2 norm.

[0029] It should be noted that the noise threshold is a custom parameter, which should be set according to image resolution, lighting conditions, etc. Taking a scene of water accumulation on an urban road as an example, you can first count the optical flow modulus distribution of the static area of ​​the road surface (the area with no obvious water flow or only slight noise fluctuations), and add 2 to 3 times the standard deviation of the mean of the distribution as the noise threshold. If you need to quickly adjust, you can also refer to the empirical value. That is, when the video resolution is 1920×1080 and the frame rate is 30fps, the noise threshold can be set to 1 to 3 pixels per frame. This setting can effectively filter out random noise caused by slight camera shake, while retaining the pixels of water that actually participate in the flow along the road, and avoiding meaningless pixels from interfering with the subsequent analysis of the flow characteristics along the road.

[0030] It should be noted that the road segment direction vector is obtained from the geometric data of the road network. Roads are usually stored in the form of centerline segments. The direction vector of the road segment can be obtained by subtracting the starting coordinate from the coordinate of the end point of the segment. The consistency coefficient along the line is used to measure the degree of consistency between the movement direction of water pixels and the direction of the road. The value ranges from 0 to 1. The larger the value, the more water pixels move along the road and the more stable the direction. The average velocity of the component along the line is used to measure the average speed of the water flow along the road.

[0031] In one embodiment of the present invention, the number of water pixels falling within the hotspot area is counted, and the percentage of water pixels in the hotspot area is obtained by dividing the number of pixels in the hotspot area. If the percentage is greater than or equal to a preset threshold, the blockage indicator is assigned a value of 1; otherwise, it is assigned a value of 0. The hotspot area is obtained by manually marking it on the road image. The preset threshold is a custom parameter. Preferably, the preset threshold is set to 0.5.

[0032] It should be noted that culverts or ditches can be selected as hotspot areas because they are key nodes for road drainage. Once blocked, rainwater can easily accumulate on the road, directly affecting road traffic capacity and even causing risks such as water accumulation and flooding. Therefore, the safety assessment of rescue routes is crucial. When the proportion of water pixels in the hotspot area is greater than or equal to the preset threshold, it indicates that there are signs of blockage in the culvert or ditch at that moment. When the proportion of water pixels is less than the preset threshold, it indicates that there is no obvious blockage at that time.

[0033] In one embodiment of the present invention, the original rainfall intensity sequence is aligned by shifting the entire sequence along the time axis by the optimal time delay, that is, the element value at time t in the original rainfall intensity sequence is mapped to the element value at time t-. For each element value, a sliding window is applied to each moment of the aligned rainfall intensity sequence. If the element value of the aligned rainfall intensity sequence at each moment within the sliding window is less than the average of the element values ​​at all moments within the sliding window, the rainfall intensity indicator is assigned a value of 1; otherwise, it is assigned a value of 0. The average of the rainfall intensity indicators at all moments within the sliding window is then calculated to obtain the lag water supply ratio. The size of the sliding window is a custom parameter. Preferably, the sliding window is set to the rounded-up value of 1 / 10 of the length of the aligned rainfall intensity sequence, and the minimum value of the sliding window size is 10.

[0034] It should be noted that the delayed water supply ratio is used to quantify the extent to which water continues to flow onto the road surface after rainfall has entered the attenuation phase, providing a quantitative basis for the subsequent calculation of the longitudinal overflow risk index.

[0035] In one embodiment of the present invention, the longitudinal overflow risk index at time i is... The calculation formula is as follows: ,in , , and Let represent the consistency coefficient along the line, the average velocity of the components along the line, the blockage indication, and the lag ratio at time i, respectively. The speed reference value is represented by g, the longitudinal slope of the road is represented by a, b, c and d, which are the first weight coefficient, the second weight coefficient, the third weight coefficient and the fourth weight coefficient respectively, and are custom parameters with a total value of 1. Preferably, a, b, c and d are set to 0.35, 0.3, 0.2 and 0.15 respectively.

[0036] It should be noted that the average slope of the road section is obtained by fitting elevation data as the longitudinal slope of the road, and in order to prevent 1 / g from being too large, its value range is limited to between 3% and 10%. The speed reference is a custom parameter, which can be the 80th to 90th percentile of the average speed of the components along the line in the past hour. This ensures that the various weighted items will not be unbalanced due to the item values ​​being too large or too small when the longitudinal overflow risk index is weighted and integrated, so as to make the risk assessment more accurate.

[0037] It should be noted that the calculation of the longitudinal overflow risk index does not involve dimensions. For example, at a certain moment, the consistency coefficient along the line, the average speed of the components along the line, the blockage indicator, and the lag water supply ratio are 0.7, 0.4, 1, and 0.6, respectively. Assuming that the speed reference is set to 0.5 and the longitudinal slope of the road is 8%, the longitudinal overflow risk index can be obtained as 0.6 according to the above calculation formula.

[0038] In one embodiment of the present invention, if the longitudinal overflow risk index is determined to be greater than or equal to a first threshold, the road area in the road image is marked as an interrupted road; if the longitudinal overflow risk index is determined to be less than the first threshold but greater than or equal to a second threshold, the road area in the road image is marked as a risky road; otherwise, it is marked as a normal road. The first threshold and the second threshold are both custom parameters. Preferably, the first threshold is set to 0.7 and the second threshold is set to 0.3.

[0039] In one embodiment of the present invention, interrupted roads are eliminated, risky roads are assigned a value of 2, and normal roads are assigned a value of 1. Alternatively, the objective function is to minimize the sum of the assigned values ​​of all roads or the longitudinal overflow risk indices based on the corresponding longitudinal overflow risk index. The optimal road rescue route from the current location of the rescue team to the accident site is obtained through a path optimization algorithm and marked in real time on a map. The path optimization algorithm can be Dijkstra's algorithm, A... Algorithms, etc., will not be elaborated here.

[0040] In one embodiment of the present invention, such as Figure 2 As shown, a road rescue route navigation system includes: The optimal time delay calculation module 201 is used to collect the rainfall intensity sequence and road image sequence at the accident site, and to calculate the water pixel ratio of the road image at each time moment to obtain the water pixel ratio sequence, and to perform correlation analysis to obtain the optimal time delay. The parameter calculation module 202 along the route is used to calculate the optical flow vector of the water pixels in the road image at each time moment, and calculate the along-route consistency coefficient and the average velocity of the along-route components by combining it with the road segment direction vector. The lag water supply ratio calculation module 203 is used to calculate the proportion of water pixels in hot spots, compare it with a preset threshold to obtain the blockage indication, align the rainfall intensity sequence according to the optimal time lag, and calculate the lag water supply ratio at each moment through a sliding window. Overflow risk index calculation module 204 is used to calculate the longitudinal overflow risk index by longitudinal slope modulation based on the consistency coefficient along the line, the average velocity of the components along the line, the blockage indication, and the lag water supply ratio. The road status marking module 205 is used to mark the status of road segments according to the longitudinal overflow risk index and simultaneously mark them on the map to obtain the road rescue route.

[0041] It should be noted that the interval and threshold sizes are set for ease of comparison. The size of the threshold depends on the amount of sample data and the base number set by those skilled in the art for each set of sample data, as long as it does not affect the proportional relationship between the parameter and the quantized value. Furthermore, the above formulas are all dimensionless calculations, and the formulas are derived from software simulations using a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0042] The embodiments of this example have been described above. However, this example is not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms based on the guidance of this example, and all of them are within the protection scope of this example.

Claims

1. A road rescue route navigation method, characterized in that, Includes the following steps: Step S101: Collect the rainfall intensity sequence and road image sequence at the accident site, and calculate the water pixel ratio of the road image at each time point to obtain the water pixel ratio sequence. Based on this, perform correlation analysis to obtain the optimal time delay. Step S102: Calculate the optical flow vector of the water pixels in the road image at each time moment, and calculate the consistency coefficient and average velocity of the road components along the route by combining it with the road segment direction vector. Step S103: Calculate the percentage of water pixels in the hotspot area, compare it with a preset threshold to obtain the blockage indication, align the rainfall intensity sequence according to the optimal time lag, and calculate the lag water supply ratio at each moment through a sliding window. Step S104: Based on the consistency coefficient along the line, the average velocity of the components along the line, the blockage indication, and the lag water supply ratio, the longitudinal overflow risk index is calculated by longitudinal slope modulation. Step S105: Mark the status of the road segment according to the longitudinal overflow risk index and mark it on the map at the same time to obtain the road rescue route.

2. The road rescue route navigation method according to claim 1, characterized in that, Using the accident site as the center point and the current location of the rescue team as the boundary point, the line connecting the two points is used as the median line. A rectangular area is generated with a fixed width using this median line. Rainfall intensity sequence is obtained by extracting grid radar rainfall information within this area. Road image sequence is obtained by using a drone to take aerial photos along the line connecting the two points to cover the area.

3. The road rescue route navigation method according to claim 1, characterized in that, First, the road region in the road image is identified by the object detection model. Then, the water region in the road image is identified by the image segmentation model. Finally, the number of water pixels falling within the road region is counted and divided by the number of pixels in the road region to obtain the proportion of water pixels in the road image.

4. The road rescue route navigation method according to claim 1, characterized in that, The difference between the element values ​​of adjacent time points in the rainfall intensity sequence and the water body pixel proportion sequence is calculated to obtain the rainfall intensity difference sequence and the water body pixel proportion difference sequence. Then, the inner product sum of the two under different lag coefficients is calculated, and the lag coefficient with the largest inner product sum is taken as the optimal time lag.

5. The road rescue route navigation method according to claim 1, characterized in that, The road image at each time step is converted into a grayscale image. The Farneback dense optical flow algorithm is used to calculate the optical flow vector of each water pixel at adjacent time steps. If the magnitude of the optical flow vector is less than the noise threshold, the water pixels in the road image at that time step are removed. Then, the remaining water pixels in the road image at each time step are statistically averaged to obtain the consistency coefficient and the average velocity of the components along the road.

6. The road rescue route navigation method according to claim 1, characterized in that, The number of water pixels falling within the hotspot area is counted, and the percentage of water pixels in the hotspot area is obtained by dividing the number of pixels in the hotspot area. If the percentage is greater than or equal to a preset threshold, the blockage indicator is set to 1; otherwise, it is set to 0. The hotspot areas are obtained by manually marking them on the road image.

7. A road rescue route navigation method according to claim 1, characterized in that, The original rainfall intensity sequence is aligned by shifting the entire sequence along the time axis by the optimal time lag. A sliding window is applied to each moment of the aligned rainfall intensity sequence. If the element value of the aligned rainfall intensity sequence at each moment within the sliding window is less than the average value of the element values ​​at all moments within the sliding window, the rainfall intensity indicator is assigned a value of 1; otherwise, it is assigned a value of 0. The average value of the rainfall intensity indicator at all moments within the sliding window is then calculated to obtain the lag water supply ratio. The size of the sliding window is a user-defined parameter.

8. A road rescue route navigation method according to claim 1, characterized in that, The minimum value of the weighted fusion result of 1 and the consistency coefficient along the route, the average speed of the component along the route, the congestion indication, and the lagging water supply ratio is taken. The four weighted terms involved in the weighted fusion are the product of the first weight coefficient and the consistency coefficient along the route, the ratio of the average speed of the component along the route to the speed reference multiplied by the second weight coefficient, 1 divided by 100 divided by the longitudinal slope of the road multiplied by the congestion indication and the third weight coefficient, and the fourth weight coefficient and the lagging water supply ratio. The first weight coefficient, the second weight coefficient, the third weight coefficient, and the fourth weight coefficient are all user-defined parameters.

9. A road rescue route navigation method according to claim 1, characterized in that, If the longitudinal overflow risk index is greater than or equal to the first threshold, the road area in the road image is marked as an interrupted road. If the longitudinal overflow risk index is less than the first threshold but greater than or equal to the second threshold, the road area in the road image is marked as a risky road. Otherwise, it is marked as a normal road. The first threshold and the second threshold are both user-defined parameters.

10. A road rescue route navigation system, characterized in that, Performing a road rescue route navigation method as described in any one of claims 1 to 9, comprising: The optimal time delay calculation module is used to collect rainfall intensity sequences and road image sequences at the accident site, and to calculate the water pixel ratio sequence of the road images at each time point to obtain the water pixel ratio sequence. Correlation analysis is then performed based on this sequence to obtain the optimal time delay. The parameter calculation module along the route is used to calculate the optical flow vector of the water pixels in the road image at each time moment, and calculate the consistency coefficient and the average velocity of the components along the route by combining it with the road segment direction vector. The lag water supply ratio calculation module is used to calculate the percentage of water pixels in hotspot areas, compare it with a preset threshold to obtain the blockage indicator, align the rainfall intensity sequence according to the optimal time lag, and calculate the lag water supply ratio at each moment through a sliding window. The overflow risk index calculation module is used to calculate the longitudinal overflow risk index by longitudinal slope modulation based on the consistency coefficient along the line, the average velocity of the components along the line, the blockage indication, and the lag supply ratio. The road status marking module is used to mark the status of road segments according to the longitudinal overflow risk index and simultaneously mark them on the map to obtain road rescue routes.

Images