A road waterlogging risk assessment method based on unmanned aerial vehicle perception

Abstract

The application discloses a road waterlogging risk assessment method based on unmanned aerial vehicle sensing, and belongs to the technical field of traffic safety risk assessment. In order to solve the problems of high-precision, sustainable road waterlogging monitoring and auxiliary decision-making, a conversion model of image pixels and actual length is constructed, a relationship model of pixel gray scale of an infrared image obtained by an infrared thermal imager and ground temperature is established, waterlogging depth is calculated by using the infrared image obtained by the infrared thermal imager, a waterlogging mask mark is generated, the area and volume of a waterlogging region are estimated, the influence of waterlogging depth on the friction between a road and a tire is considered, the friction coefficient under a wet and slippery condition is calculated, the friction coefficient under the wet and slippery condition is corrected by introducing shear stress to obtain the friction coefficient of shear effect, then the recommended vehicle speed is calculated based on the current friction level of the road and the change rate of waterlogging, and a traffic comprehensive risk score model of the road under the waterlogging state is established to assess the road waterlogging risk.

Description

Technical Field

[0001] This invention belongs to the field of traffic safety risk assessment technology, specifically relating to a method for assessing road water accumulation risk based on UAV perception. Background Technology

[0002] Road flooding often causes traffic congestion and even poses accident risks, impacting citizen travel efficiency and public safety. Timely monitoring of road flooding information is an indispensable part of ensuring urban operation. However, current road flooding detection relies on sensors deployed at fixed locations or manual patrols, resulting in slow response times, numerous blind spots, and difficulty in covering the complex urban road structure. Furthermore, flooding phenomena are characterized by rapid, short-term changes and wide spatial distribution, making it difficult for traditional methods to detect quickly and effectively. There is an urgent need to develop a comprehensive method integrating image acquisition, information analysis, and deductive judgment to overcome the current bottlenecks in flooding monitoring. Summary of the Invention

[0003] The problem this invention aims to solve is to achieve high-precision and sustainable monitoring and decision support for road waterlogging, and proposes a road waterlogging risk assessment method based on UAV perception.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A method for assessing road flooding risk based on UAV perception includes the following steps:

[0006] S1. The UAV is equipped with an infrared thermal imager and a visible light camera to collect road images and construct a conversion model between image pixels and actual length;

[0007] S2. Establish a model of the relationship between pixel grayscale of infrared images obtained by infrared thermal imagers and surface temperature, which can be used to determine water accumulation at night, in low visibility or complex weather conditions;

[0008] S3. Calculate the water depth using infrared images obtained from an infrared thermal imager;

[0009] S4. The infrared and visible light images simultaneously acquired by the UAV are fused together to generate water accumulation mask markers;

[0010] S5. Estimate the area and volume of the water accumulation area based on the water accumulation mask markers obtained in step S4;

[0011] S6. Considering the effect of water depth on the friction between the road and tires, calculate the coefficient of friction under wet and slippery conditions;

[0012] S7. Shear stress is introduced to correct the friction coefficient under wet and slippery conditions to obtain the friction coefficient of the shear effect. Then, the recommended vehicle speed is calculated based on the current friction level of the road and the rate of change of the water accumulation.

[0013] S8. Establish a comprehensive traffic risk scoring model for roads under waterlogging conditions. Then, introduce the temporal intensity change rate of UAV images and external real-time meteorological rainfall intensity information to correct the comprehensive traffic risk scoring model, and obtain the corrected comprehensive traffic risk scoring model to assess the risk of road waterlogging.

[0014] Furthermore, step S1 constructs a conversion model between image pixels and ground length based on the drone's flight altitude, the camera's horizontal field of view, and the number of horizontal pixels in the image. The expression is:

[0015]

[0016] in, The actual length of the ground corresponding to each pixel; The drone's flight altitude is obtained from the drone monitoring platform. This is the horizontal field of view of the camera, obtained from the camera parameters; This represents the number of horizontal pixels in the image, obtained from the image resolution information.

[0017] Furthermore, the expression for the surface temperature obtained in step S2 based on the infrared pixel grayscale is as follows:

[0018]

[0019] in, Let be the surface temperature at coordinates (i, j); The pixel grayscale value of the infrared image at coordinates (i, j) is obtained directly from the acquired image; , , These are the temperature conversion gain coefficient, grayscale shift coefficient, and temperature saturation adjustment coefficient, respectively.

[0020] Furthermore, the formula for calculating the water depth in step S3 is as follows:

[0021]

[0022] in, Let (i, j) be the depth of the water accumulation at coordinate (i, j). Ground reflectivity; , , These are the temperature-depth conversion reference coefficient, temperature-depth attenuation index, and reflectivity influence factor, respectively.

[0023] Furthermore, step S4 relies on the ability of the drone to simultaneously acquire infrared and visible light images, fusing the two types of sensing signals, as expressed in the following expression:

[0024]

[0025]

[0026] in, The saliency fusion response value at coordinate (i, j); The visible light image gray level at coordinates (i, j) is obtained from the acquired image; Let (i, j) be the Laplacian operator for the grayscale image at coordinates (i, j); To determine the threshold; Mark the water accumulation mask at coordinates (i, j). , These are the temperature adjustment factor and the texture adjustment factor, respectively, and are also used to adjust the dimensions.

[0027] Furthermore, the specific implementation method of step S5 includes the following steps:

[0028] S5.1. Calculate the area A of the water accumulation region using the water accumulation mask markers and the actual ground length corresponding to each pixel;

[0029] S5.2. Calculate the volume of the water accumulation area using water accumulation mask markers and water depth. .

[0030] Furthermore, the formula for calculating the coefficient of friction under wet and slippery conditions in step S6 is as follows:

[0031]

[0032] in, Let be the coefficient of friction at coordinate (i, j) under wet and slippery conditions; The reference friction coefficient for dry road surfaces is determined from the vehicle manual or by testing. The density of water; The tread water-resistant factor is determined by the vehicle manual or testing. Tire pressure, determined from the vehicle manual or by testing; It is a nonlinear friction variation factor; It serves as a pressure adjustment factor and is also used to adjust the dimensions.

[0033] Furthermore, the specific implementation method of step S7 includes the following steps:

[0034] S7.1. Establish the friction coefficient for the shear effect, expressed as follows:

[0035]

[0036] in, Let be the friction coefficient at coordinate (i, j) considering the shear effect; This is a correction factor; The shear stress at coordinate (i, j) is obtained through dynamic modeling and simulation. The boundary shear force of the tire surface is determined by testing or provided in the material handbook.

[0037] S7.2. Calculate the recommended vehicle speed. The calculation process is as follows:

[0038]

[0039]

[0040] in, Recommended speed; It is the acceleration due to gravity; Braking distance, obtained from vehicle parameters; It is used as an adjustment factor and also for adjusting the dimensions; The average friction coefficient for the road area; The velocity decay rate; This is the water quantity influence coefficient, and it is also used to adjust the dimensions. The incremental volume change is calculated by subtracting the volumes at different times.

[0041] Furthermore, the specific implementation method of step S8 includes the following steps:

[0042] S8.1. The relationship between water volume, area, and road width, as well as the average friction coefficient of the road area and the current water growth rate, are integrated. Then, after normalization and weighted summation, a comprehensive traffic risk score is generated, expressed as:

[0043]

[0044] Where R represents the overall risk score; The road width is obtained from the design documents; To and The corresponding volume change time is determined by technical personnel based on actual needs; , These are the depth normalization coefficient and the rate of change normalization coefficient, respectively; , , These are respectively the water depth weight, friction risk weight, and dynamic change weight, which are also used to adjust the dimensions.

[0045] S8.2. The comprehensive risk score obtained in step S8.1 is corrected to obtain the corrected comprehensive traffic risk score, expressed as:

[0046]

[0047]

[0048] in, This is a score adjustment value; The rate of change of image intensity; Real-time rainfall intensity, obtained from the meteorological platform; , These are the image variation standardization factor and the rainfall intensity standardization factor, respectively. , These are image change weights and rainfall weights, respectively. This is the revised comprehensive traffic risk score.

[0049] The beneficial effects of this invention are:

[0050] The present invention discloses a road flood risk assessment method based on UAV perception. It uses infrared and visible light image fusion to improve the accuracy of identification, uses temperature and physical parameters to perform inversion calculation to calculate the depth of floodwater, and corrects the risk trend by changing the image and meteorological information. It can quickly, extensively and precisely assess the impact of floodwater on road traffic risk, and provide technical support for dynamic safety management in various road environments.

[0051] The present invention discloses a road flooding risk assessment method based on UAV perception, which accurately identifies road flooding areas from image information, quantifies the depth and area of ​​the flooding, and then analyzes its impact on vehicle traffic, such as friction loss, speed recommendations, and risk assessment, so as to achieve non-contact, high-precision, and sustainable road flooding monitoring and auxiliary decision-making. Attached Figure Description

[0052] Figure 1 This is a flowchart of a road flooding risk assessment method based on UAV perception, as described in this invention.

[0053] Figure 2 This is a graph showing the friction coefficient, recommended vehicle speed, and comprehensive risk score calculated in an embodiment of the present invention. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the invention and are not intended to limit the invention; that is, the described specific embodiments are merely a part of the embodiments of the invention, and not all of them. The components of the specific embodiments of the invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations, and the invention may also have other embodiments.

[0055] Therefore, the following detailed description of specific embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected specific embodiments of the invention. All other specific embodiments obtained by those skilled in the art based on these specific embodiments without inventive effort are within the scope of protection of this invention.

[0056] To further understand the invention's content, features, and effects, the following specific embodiments are provided, along with accompanying drawings. Figure 1 and attached Figure 2 Detailed explanation is as follows:

[0057] Example 1:

[0058] A method for assessing road flooding risk based on UAV perception includes the following steps:

[0059] S1. The UAV is equipped with an infrared thermal imager and a visible light camera to collect road images and construct a conversion model between image pixels and actual length;

[0060] Furthermore, step S1 constructs a conversion model between image pixels and ground length based on the drone's flight altitude, the camera's horizontal field of view, and the number of horizontal pixels in the image. The expression is:

[0061]

[0062] in, The actual length of the ground corresponding to each pixel; The drone's flight altitude is obtained from the drone monitoring platform. This is the horizontal field of view of the camera, obtained from the camera parameters; This represents the number of horizontal pixels in the image, obtained from the image resolution information.

[0063] Furthermore, drones are equipped with infrared thermal imagers and visible light cameras to assist in dynamic road monitoring. Before conducting water accumulation analysis, a key task is to accurately map these pixel coordinates to actual physical distances on the ground. This step relies on the drone's flight altitude, viewing angle, and the number of horizontal pixels in the image to construct an image conversion formula and derive a scaling factor, which can convert each image pixel into a corresponding unit length on the ground.

[0064] S2. Establish a model of the relationship between pixel grayscale of infrared images obtained by infrared thermal imagers and surface temperature, which can be used to determine water accumulation at night, in low visibility or complex weather conditions;

[0065] Furthermore, the expression for the surface temperature obtained in step S2 based on the infrared pixel grayscale is as follows:

[0066]

[0067] in, Let be the surface temperature at coordinates (i, j); The pixel grayscale value of the infrared image at coordinates (i, j) is obtained directly from the acquired image; , , These are the temperature conversion gain coefficient, grayscale shift coefficient, and temperature saturation adjustment coefficient, respectively.

[0068] Furthermore, the infrared thermal imager carried by the drone can quickly acquire images of the surface thermal radiation of the covered area at a certain flight altitude. The grayscale values ​​in the infrared images reflect the thermal response intensity of each area, but they do not have direct physical meaning. By establishing a nonlinear function model between infrared grayscale and thermal temperature, and combining it with the image data collected by the drone, the grayscale can be converted into the base surface temperature. This feature is applicable to the assessment of water accumulation at night, in low visibility, or in complex weather conditions, enabling the drone to continuously and stably perform monitoring tasks even in harsh environments.

[0069] S3. Calculate the water depth using infrared images obtained from an infrared thermal imager;

[0070] Furthermore, the formula for calculating the water depth in step S3 is as follows:

[0071]

[0072] in, Let (i, j) be the depth of the water accumulation at coordinate (i, j). Ground reflectivity; , , These are the temperature-depth conversion reference coefficient, temperature-depth attenuation index, and reflectivity influence factor, respectively.

[0073] Furthermore, the surface temperature recorded in infrared images is affected by many factors, with water accumulation having the most significant effect on suppressing heat radiation emitted from the road surface. By using high-altitude infrared images acquired by UAVs, combined with surface material parameters, the distribution of water depth can be deduced by observing changes in the temperature field. This step constructs a model related to thermophysics, using the temperature field acquired by the UAV as input, and combining it with reflection parameters to calculate the water depth at each water accumulation point.

[0074] S4. The infrared and visible light images simultaneously acquired by the UAV are fused together to generate water accumulation mask markers;

[0075] Furthermore, step S4 relies on the ability of the drone to simultaneously acquire infrared and visible light images, fusing the two types of sensing signals, as expressed in the following expression:

[0076]

[0077]

[0078] in, The saliency fusion response value at coordinate (i, j); The visible light image gray level at coordinates (i, j) is obtained from the acquired image; Let (i, j) be the Laplacian operator for the grayscale image at coordinates (i, j); To determine the threshold; Mark the water accumulation mask at coordinates (i, j). , These are the temperature adjustment factor and the texture adjustment factor, respectively, and are also used to adjust the dimensions.

[0079] Furthermore, leveraging the ability of drones to simultaneously acquire infrared and visible light images, the two types of sensing signals can be fused to improve the accuracy of water accumulation identification. A single infrared thermal image may lead to misjudgments due to road surface interference. Introducing texture information from visible light images for edge enhancement processing, and utilizing the drone's simultaneous acquisition trajectory, allows for pixel-level alignment of the two images. Image fusion can filter out false water accumulation areas, resulting in more accurate identification. By setting temperature and texture weight functions and performing threshold segmentation, a precise water accumulation mask can be generated, providing a data foundation for subsequent area and volume calculations.

[0080] S5. Estimate the area and volume of the water accumulation area based on the water accumulation mask markers obtained in step S4;

[0081] Furthermore, the specific implementation method of step S5 includes the following steps:

[0082] S5.1. Using the water accumulation mask markers and the actual ground length corresponding to each pixel, calculate the area A of the water accumulation region. The calculation formula is as follows:

[0083]

[0084] Where A is the total area covered by the floodwater;

[0085] Furthermore, after the drone image analysis identifies the mask of the water accumulation area, the total water coverage area is accurately calculated by using the actual length of the ground corresponding to each pixel. Drones have advantages such as a wide field of view and high resolution, and can quickly and accurately segment water accumulation areas even in complex road structures such as municipal environments, overpasses, and elevated roads.

[0086] S5.2. Calculate the volume of the water accumulation area using water accumulation mask markers and water depth. The calculation formula is:

[0087]

[0088] in, This represents the total volume of accumulated water.

[0089] Furthermore, by using the depth and area of ​​the flooded area, the total volume of water in the road area can be calculated. Compared with traditional point-based monitoring methods, drones can cover a wider area in a shorter time. Estimating the total amount of water without contact can also support subsequent dynamic monitoring and trend prediction of the entire area.

[0090] S6. Considering the effect of water depth on the friction between the road and tires, calculate the coefficient of friction under wet and slippery conditions;

[0091] Furthermore, the formula for calculating the coefficient of friction under wet and slippery conditions in step S6 is as follows:

[0092]

[0093] in, Let be the coefficient of friction at coordinate (i, j) under wet and slippery conditions; The reference friction coefficient for dry road surfaces is determined from the vehicle manual or by testing. The density of water; The tread water-resistant factor is determined by the vehicle manual or testing. Tire pressure, determined from the vehicle manual or by testing; It is a nonlinear friction variation factor; It serves as a pressure adjustment factor and is also used to adjust the dimensions.

[0094] Furthermore, the images acquired by drones are not only used to identify water accumulation, but can also indirectly reveal the traffic hazards caused by changes in water depth. As water deepens, the effective friction between the road and tires decreases rapidly, easily leading to traffic accidents. This model uses water depth as input and establishes a friction decay function that considers the effect of water adhesion pressure, dynamically representing the decreasing trend of physical friction under slippery conditions. Combined with the drone's area monitoring range, this model can analyze which areas within a complete road segment have higher risks, achieving image-based risk mapping.

[0095] S7. Shear stress is introduced to correct the friction coefficient under wet and slippery conditions to obtain the friction coefficient of the shear effect. Then, the recommended vehicle speed is calculated based on the current friction level of the road and the rate of change of the water accumulation.

[0096] Furthermore, the specific implementation method of step S7 includes the following steps:

[0097] S7.1. Establish the friction coefficient for the shear effect, expressed as follows:

[0098]

[0099] in, Let be the friction coefficient at coordinate (i, j) considering the shear effect; This is a correction factor; The shear stress at coordinate (i, j) is obtained through dynamic modeling and simulation. The boundary shear force of the tire surface is determined by testing or provided in the material handbook.

[0100] Furthermore, accurately simulating the risk of loss of control on slippery roads requires more than just water depth. During emergency steering and braking, vehicles experience high shear stress, and insufficient lateral friction can easily lead to skidding. This step uses water depth measurements obtained from UAV images and maps it into a nonlinear friction model. A mechanical model is also introduced to adjust the basic friction, making the model more closely reflect the complex contact process between vehicles and roads in reality. UAVs can perform large-scale, multi-point friction estimations, providing more reliable simulation support for vehicle dynamics and lane guidance.

[0101] S7.2. Calculate the recommended vehicle speed. The calculation process is as follows:

[0102]

[0103]

[0104] in, Recommended speed; It is the acceleration due to gravity; Braking distance, obtained from vehicle parameters; It is used as an adjustment factor and also for adjusting the dimensions; The average friction coefficient for the road area; The velocity decay rate; This is the water quantity influence coefficient, and it is also used to adjust the dimensions. The incremental volume change is calculated by subtracting the volumes at different times.

[0105] Furthermore, the depth of the water can be analyzed from the images acquired by the drone, and the friction level can be calculated. This step is used to calculate the recommended driving speed. The model is based on the current friction level of the road and the rate of change of the water. After estimating the rate of water growth by using the difference between different time-series images, it is combined with the vehicle's physical braking distance to output a current safe driving speed recommendation.

[0106] S8. Establish a comprehensive traffic risk scoring model for roads under waterlogging conditions. Then, introduce the temporal intensity change rate of UAV images and external real-time meteorological rainfall intensity information to correct the comprehensive traffic risk scoring model, and obtain the corrected comprehensive traffic risk scoring model to assess the risk of road waterlogging.

[0107] Furthermore, the specific implementation method of step S8 includes the following steps:

[0108] S8.1. The relationship between water volume, area, and road width, as well as the average friction coefficient of the road area and the current water growth rate, are integrated. Then, after normalization and weighted summation, a comprehensive traffic risk score is generated, expressed as:

[0109]

[0110] Where R represents the comprehensive traffic risk score; The road width is obtained from the design documents; To and The corresponding volume change time is determined by technical personnel based on actual needs; , These are the depth normalization coefficient and the rate of change normalization coefficient, respectively; , , These are respectively the water depth weight, friction risk weight, and dynamic change weight, which are also used to adjust the dimensions.

[0111] Furthermore, based on the results of UAV image analysis, this step establishes a risk scoring function that can more comprehensively represent the traffic risk level of roads under current flooding conditions. This function integrates the relationships between volume, area, and road width, as well as the average friction coefficient and the current water growth rate, and then, through normalization and weighted summation, generates a comprehensive traffic risk score.

[0112] S8.2. Correct the comprehensive traffic risk score obtained in step S8.1 to obtain the corrected comprehensive traffic risk score, expressed as:

[0113]

[0114]

[0115] in, This is a score adjustment value; The rate of change of image intensity; Real-time rainfall intensity, obtained from the meteorological platform; , These are the image variation standardization factor and the rainfall intensity standardization factor, respectively. , These are image change weights and rainfall weights, respectively. This is the revised comprehensive traffic risk score.

[0116] Furthermore, in addition to static risk factors, the trend of short-term risk changes should also be considered when conducting comprehensive risk scoring. To this end, this step introduces the temporal intensity change rate of UAV images and combines it with real-time external meteorological rainfall intensity information to form a trend correction model. The temporal change rate of images can reflect the trend of water accumulation expansion, while rainfall intensity shows the external water replenishment capacity. Combining the two into a dynamic correction term can improve the scoring model's ability to perceive risk factors such as sudden weather and continuous heavy rain.

[0117] By employing the above methods, the image information acquired by drones can be transformed into the actual depth, area, and volume of surface water. Combined with friction theory and rainfall variations, this information can be used to calculate recommended road speeds and traffic risk scores. This approach is adaptable to different weather and environmental conditions, providing quantitative data for road safety and supporting urban drainage scheduling, traffic management, and emergency preparedness.

[0118] The application examples of this embodiment are illustrated below:

[0119] A drone was used to analyze the water accumulation on a road. The drone flew at an altitude of 15m and was equipped with an infrared and a high-definition camera. The road surface material was asphalt with a reflectivity of 0.12, and the vehicle tire pressure was 200 kPa. Using this method, the friction coefficient considering shear effects under wet and slippery conditions was calculated to be 0.74. The recommended speed for this road was 60 km / h, and the corrected comprehensive risk score was 0.75.

[0120] It should be noted that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0121] Although this application has been described above with reference to specific embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of this application. In particular, as long as there is no structural conflict, the features in the specific embodiments disclosed in this application can be combined with each other in any way. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, this application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for assessing road flooding risk based on UAV perception, characterized in that, Includes the following steps: S1. The UAV is equipped with an infrared thermal imager and a visible light camera to collect road images and construct a conversion model between image pixels and actual length; S2. Establish a model of the relationship between pixel grayscale of infrared images obtained by infrared thermal imagers and surface temperature, which can be used to determine water accumulation at night, in low visibility or complex weather conditions; The expression for the surface temperature obtained in step S2 based on infrared pixel grayscale is: ； in, Let be the surface temperature at coordinates (i, j); The pixel grayscale value of the infrared image at coordinates (i, j) is obtained directly from the acquired image; , , These are the temperature conversion gain coefficient, grayscale offset coefficient, and temperature saturation adjustment coefficient, respectively. S3. Calculate the water depth using infrared images obtained from an infrared thermal imager; The formula for calculating the water depth in step S3 is: ； in, Let (i, j) be the depth of the water accumulation at coordinate (i, j). Ground reflectivity; , , These are the temperature-depth conversion reference coefficient, temperature-depth attenuation index, and reflectivity influence factor, respectively. S4. The infrared and visible light images simultaneously acquired by the UAV are fused together to generate water accumulation mask markers; S5. Estimate the area and volume of the water accumulation area based on the water accumulation mask markers obtained in step S4; S6. Considering the effect of water depth on the friction between the road and tires, calculate the coefficient of friction under wet and slippery conditions; The formula for calculating the coefficient of friction under wet and slippery conditions in step S6 is: ； in, Let be the coefficient of friction at coordinate (i, j) under wet and slippery conditions; The reference friction coefficient for dry road surfaces is determined from the vehicle manual or by testing. The density of water; The tread water-resistant factor is determined by the vehicle manual or testing. Tire pressure, determined from the vehicle manual or by testing; It is a nonlinear friction variation factor; It serves as a pressure adjustment factor and is also used to adjust the dimensions. S7. Shear stress is introduced to correct the friction coefficient under wet and slippery conditions to obtain the friction coefficient of the shear effect. Then, the recommended vehicle speed is calculated based on the current friction level of the road and the rate of change of the water accumulation. The specific implementation method of step S7 includes the following steps: S7.

1. Establish the friction coefficient for the shear effect, expressed as follows: ； in, Let be the friction coefficient at coordinate (i, j) considering the shear effect; This is a correction factor; The shear stress at coordinate (i, j) is obtained through dynamic modeling and simulation. The boundary shear force of the tire surface is determined by testing or provided in the material handbook. S7.

2. Calculate the recommended vehicle speed. The calculation process is as follows: ； ； in, Recommended speed; It is the acceleration due to gravity; Braking distance, obtained from vehicle parameters; It serves as an adjustment factor and is also used to adjust the dimensions; The average friction coefficient for the road area; The velocity decay rate; This is the water quantity influence coefficient, and it is also used to adjust the dimensions. The incremental volume change is calculated by the difference between the volumes at different times; S8. Establish a comprehensive traffic risk scoring model for roads under waterlogging conditions. Then, introduce the temporal intensity change rate of UAV images and external real-time meteorological rainfall intensity information to correct the comprehensive traffic risk scoring model, and obtain the corrected comprehensive traffic risk scoring model to assess the risk of road waterlogging.

2. The road flooding risk assessment method based on UAV perception according to claim 1, characterized in that, Step S1 uses the drone's flight altitude, the camera's horizontal field of view, and the number of horizontal pixels in the image to construct a conversion model between image pixels and ground length, expressed as: ； in, The actual length of the ground corresponding to each pixel; The drone's flight altitude is obtained from the drone monitoring platform. This is the horizontal field of view of the camera, obtained from the camera parameters; This represents the number of horizontal pixels in the image, obtained from the image resolution information.

3. The road flooding risk assessment method based on UAV perception according to claim 2, characterized in that, Step S4 relies on the ability of the drone to simultaneously acquire infrared and visible light images, fusing the two types of sensing signals, as expressed in the following expression: ； ； in, The saliency fusion response value at coordinate (i, j); The visible light image gray level at coordinates (i, j) is obtained from the acquired image; Let (i, j) be the Laplacian operator for the grayscale image at coordinates (i, j); To determine the threshold; Mark the water accumulation mask at coordinates (i, j). , These are the temperature adjustment factor and the texture adjustment factor, respectively, and are also used to adjust the dimensions.

4. The road flooding risk assessment method based on UAV perception according to claim 3, characterized in that, The specific implementation method of step S5 includes the following steps: S5.

1. Calculate the area A of the water accumulation region using the water accumulation mask markers and the actual ground length corresponding to each pixel; S5.

2. Calculate the volume of the water accumulation area using water accumulation mask markers and water depth. .

5. The road flooding risk assessment method based on UAV perception according to claim 4, characterized in that, The specific implementation method of step S8 includes the following steps: S8.

1. The relationship between water volume, area, and road width, as well as the average friction coefficient of the road area and the current water growth rate, are integrated. Then, after normalization and weighted summation, a comprehensive traffic risk score is generated, expressed as: ； Where R represents the overall risk score; The road width is obtained from the design documents; To and The corresponding volume change time is determined by technical personnel based on actual needs; , These are the depth normalization coefficient and the rate of change normalization coefficient, respectively; , , These are respectively the water depth weight, friction risk weight, and dynamic change weight, which are also used to adjust the dimensions. S8.

2. The comprehensive risk score obtained in step S8.1 is corrected to obtain the corrected comprehensive traffic risk score, expressed as: ； ； in, This is a score adjustment value; The rate of change of image intensity; Real-time rainfall intensity, obtained from the meteorological platform; , These are the image variation standardization factor and the rainfall intensity standardization factor, respectively. , These are image change weights and rainfall weights, respectively. This is the revised comprehensive traffic risk score.

Images