Intelligent assessment method for water damage risk of dispersive soil channel

Abstract

The invention discloses a dispersive soil channel water damage risk intelligent assessment method, and relates to the technical field of water conservancy projects. According to the dispersive soil channel water damage risk intelligent assessment method provided by the invention, before dispersive soil channel water damage, the causal model and the deep learning network are utilized to accurately and quickly determine the position where water damage easily occurs in the dispersive soil channel, and the risk degree is determined according to the grading standard, so that protection measures can be reasonably designed in engineering design, and the risk assessment efficiency is improved. The water damage risk is reduced or avoided, and the damage of water damage to the channel is effectively prevented; in the on-site inspection process in the operation period, key inspection positions can be set according to the identified and determined positions prone to water damage, the water damage positions can be found as soon as possible, the situation that the water damage degree is increased due to omission is avoided, and the engineering repair difficulty and cost are reduced; the method has important engineering significance for dispersive soil channel engineering design, safety evaluation, hazard prevention, emergency rescue and the like.

Description

Technical Field

[0001] The invention relates to the technical field of water conservancy projects, and in particular to an intelligent assessment method for water damage risks of decentralized earth channels. Background Art

[0002] Dispersive soil is a special type of soil that is water-sensitive. In pure water, the cohesive structure of clay particles is destroyed by itself and dispersed into original clay particles. Dispersive soil has a low ability to resist pure water penetration and damage. Dispersive soil channels are prone to water damage due to soil dispersion, hydraulic erosion, etc., thus inducing accidents.

[0003] Dispersed soil channels are easily eroded by the confluence outside the channel and cause water damage accidents. The location, speed, degree of damage, etc. of the erosion development of dispersed soil channels are affected by many factors, such as the channel material, density, slope and other internal conditions, external conditions such as confluence area and precipitation, and the contact relationship between internal and external conditions. Dispersed soil channels are long. At present, due to the inability to accurately determine the location prone to water damage and its risks, the entire length of the channel is often treated indiscriminately during the design of dispersed soil channels or the safety inspection of built dispersed soil channels, resulting in the lack of key risk locations in the design process, and the lack of focus on risk-prone areas during on-site inspections, and the failure to eliminate risks early in the design stage and inspections. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides an intelligent assessment method for water damage risk of dispersed earth channels, which solves the problem that there is no method for water damage risk assessment for dispersed earth channels.

[0005] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0006] A method for intelligently assessing the water damage risk of decentralized earth channels is provided, which comprises the following steps:

[0007] S1. Use drones to inspect the dispersed soil channels to be evaluated and obtain the corresponding channel image data in real time;

[0008] S2, extracting features from the channel image data to obtain channel information and terrain information;

[0009] S3, obtaining the operation information of the dispersed soil channel to be evaluated;

[0010] S4, performing data processing on the channel information, the terrain information, and the operation information to obtain processed channel information, terrain information, and operation information;

[0011] S5. Construct an intelligent identification model for water damage risks;

[0012] S6. Input the processed channel information, terrain information and operation information into the water damage risk intelligent identification model to obtain the central location of the water damage prone area and its water damage information;

[0013] S7. Calculate the actual location of the water-prone area based on the spatial coordinates, flight attitude, and center location of the water-prone area of ​​the dispersed earth channel;

[0014] S8. Combining the processed operation information, terrain information, water damage information of water damage prone areas, and actual locations, conduct risk assessment on water damage prone areas to obtain corresponding risk assessment results.

[0015] Furthermore, step S1 includes the following steps:

[0016] S1-1. Use drones to inspect the dispersed soil channels to be evaluated, collect the corresponding initial channel image data and flight environment data, and upload them to the server in real time;

[0017] S1-2. Based on the flight environment data, a data processing method is selected to perform image enhancement and geometric radiation correction on the initial channel image data to obtain processed channel image data, namely, channel image data.

[0018] Furthermore, channel information includes channel height and slope ratio; terrain information includes catchment area and initial height difference between the channel top and outside the channel; operation information includes channel operation time, precipitation intensity and rainfall duration; water damage information includes water damage depth and water damage width at the water damage location.

[0019] Furthermore, the data processing in step S4 includes data correction and data cleaning, wherein:

[0020] The specific process of data correction is as follows:

[0021] Performing position correction, time correction and unit correction on the channel information, terrain information and operation information to obtain corrected channel information, terrain information and operation information;

[0022] The specific process of data cleaning is as follows:

[0023] The corrected channel information, terrain information and operation information are screened, abnormal data are deleted, and missing data are completed using empirical values ​​or interpolation methods. The completed data and normal data are retained to obtain the processed channel information, terrain information and operation information.

[0024] Furthermore, the water damage risk intelligent identification model in step S5 adopts a deep learning network based on a causal model, including an input layer, multiple hidden layers and an output layer, and each hidden layer includes multiple hidden units.

[0025] Furthermore, the training process of the water damage risk intelligent identification model in step S5 includes the following steps:

[0026] S5-1, obtaining the training water damage position, training water damage information, training channel information, training terrain information and training operation information of the dispersed earth channel that has suffered water damage;

[0027] S5-2, constructing an initial decentralized earth channel water damage training database based on training water damage channel information, training terrain information and training operation information;

[0028] S5-3, performing data cleaning on all training data in the initial decentralized earth canal risk training database to obtain a decentralized earth canal water damage training database;

[0029] S5-4, using the training water damage location and training water damage information as labels of the water damage risk intelligent identification model;

[0030] S5-5. Input the training channel information, training terrain information, training operation information and labels in the decentralized earth channel water damage training database into the water damage risk intelligent identification model for training, adjust the number of hidden layers, the number of hidden units and the weight parameters of the water damage risk intelligent identification model, and obtain the trained water damage risk intelligent identification model.

[0031] Further, step S5-1 includes the following steps:

[0032] S5-1-1. Use drones to scan dispersed earth channels that have been damaged by water to determine the location of training water damage and its initial training terrain information;

[0033] S5-1-2. Combined with manual auxiliary marking, use discontinuity analysis technology and SURF algorithm to map the training water damage location, and extract the initial training water damage information and initial training channel information of the training water damage location;

[0034] S5-1-3, based on the initial training terrain information, the initial training water damage information, and the initial training channel information, the initial training operation information of the training water damage location is determined based on the geographical location;

[0035] S5-1-4, uniformly correct the position, time and unit of the initial training terrain information, the initial training water damage information, the initial training channel information and the initial training operation information to obtain the training water damage channel information, the training terrain information and the training operation information.

[0036] Furthermore, the specific process of step S6 is:

[0037] The channel information, terrain information, and operation information of the dispersed earth channels to be evaluated are matched with the channel information, terrain information, and operation information of the dispersed earth channels that have suffered water damage through the intelligent identification model of water damage risk. The data consistent with the dispersed earth channels that have suffered water damage are regarded as areas prone to water damage, and the corresponding central location and its water damage information are output.

[0038] Furthermore, the calculation formula of step S7 is:

[0039] H=f×GSD / a

[0040] X=X0+H×tan(α)×cos(β)

[0041] Y=Y0+H×tan(α)×sin(β)

[0042] Among them, X and Y represent the actual position of the water-prone area, X0 and Y0 represent the spatial coordinates of the UAV, H represents the vertical distance from the UAV to the canal top at the center of the water-prone area, f represents the focal length, GSD represents the ground resolution, a represents the pixel size, α represents the angle between the center of the water-prone area and the UAV and the vertical, β represents the angle between the center of the water-prone area and the horizontal axis in the UAV spatial coordinates, tan(·), cos(·), and sin(·) represent the tangent function, cosine function, and sine function, respectively.

[0043] Further, step S8 includes the following steps:

[0044] S8-1. Based on the depth and width of water damage in each water damage prone area, each water damage prone area is classified to obtain the corresponding water damage level; the water damage level is divided into low risk, medium risk, high risk and extremely dangerous;

[0045] S8-2. Based on the catchment area classification conditions, the dispersed soil channels to be evaluated are classified to obtain the corresponding catchment levels; the catchment levels are divided into low risk, medium risk, high risk and extremely dangerous;

[0046] S8-3. Based on the classification conditions of precipitation intensity and rainfall duration, the dispersed soil channels to be evaluated are classified to obtain the corresponding rainfall levels; wherein the rainfall levels are divided into low risk, medium risk, high risk and extremely dangerous;

[0047] S8-4. Determine the risk assessment results based on the water damage level, runoff level and rainfall level.

[0048] The beneficial effects of the present invention are as follows: a method for intelligently assessing the water damage risk of decentralized earth channels is proposed. Before the decentralized earth channels are damaged by water, the causal model and the deep learning network are used to accurately and quickly determine the locations in the decentralized earth channels that are prone to water damage, and the risk level is determined according to the classification standards. Therefore, protective measures can be reasonably designed in engineering design to reduce or avoid the occurrence of water damage risks and effectively prevent water damage to channels. In addition, during the on-site inspection during operation, key inspection locations can be set according to the identified water-prone locations, so that the water damage locations can be discovered early to avoid the increase of the degree of water damage due to omissions, thereby reducing the difficulty and cost of engineering repair. This has important engineering significance for decentralized earth channel engineering design, safety assessment, hazard prevention and emergency rescue. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a method flow chart;

[0050] Figure 2 This is a schematic diagram of the water damage channel information at the water damage location. DETAILED DESCRIPTION

[0051] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0052] like Figure 1 As shown, a decentralized earth channel water damage risk intelligent assessment method includes the following steps:

[0053] S1. Use drones to inspect the dispersed soil channels to be evaluated and obtain the corresponding channel image data in real time;

[0054] Step S1 includes the following steps:

[0055] S1-1. Use drones to inspect the dispersed soil channels to be evaluated, collect the corresponding initial channel image data and flight environment data, and upload them to the server in real time;

[0056] S1-2. Based on the flight environment data, a data processing method is selected to perform image enhancement and geometric radiation correction on the initial channel image data to obtain processed channel image data, namely, channel image data.

[0057] S2, extracting features from the channel image data to obtain channel information and terrain information;

[0058] Channel information includes channel height and slope ratio; terrain information includes catchment area and initial height difference between the top of the channel and outside the channel;

[0059] Feature extraction can be performed through existing image processing methods or deep learning networks.

[0060] S3, obtaining the operation information of the dispersed soil channel to be evaluated;

[0061] Operational information includes channel operation time, precipitation intensity, and rainfall duration;

[0062] S4, performing data processing on the channel information, the terrain information, and the operation information to obtain processed channel information, terrain information, and operation information;

[0063] The data processing in step S4 includes data correction and data cleaning, wherein:

[0064] The specific process of data correction is as follows:

[0065] Performing position correction, time correction and unit correction on the channel information, terrain information and operation information to obtain corrected channel information, terrain information and operation information;

[0066] The specific process of data cleaning is as follows:

[0067] The corrected channel information, terrain information and operation information are screened, abnormal data are deleted, and missing data are completed using empirical values ​​or interpolation methods. The completed data and normal data are retained to obtain the processed channel information, terrain information and operation information.

[0068] During the screening process, the corrected channel information, terrain information and operation information are processed using a clustering method, a scatter plot or a normal distribution plot to obtain corresponding cluster plots, scatter plots and normal distribution plots, and outliers in the plots are removed while other data are retained.

[0069] S5. Construct an intelligent identification model for water damage risks;

[0070] The water damage risk intelligent identification model in step S5 adopts a deep learning network based on a causal model, including an input layer, multiple hidden layers and an output layer, each hidden layer including multiple hidden units. The training process of the water damage risk intelligent identification model in step S5 includes the following steps:

[0071] S5-1, obtaining the training water damage position, training water damage information, training channel information, training terrain information and training operation information of the dispersed earth channel that has suffered water damage;

[0072] Step S5-1 includes the following steps:

[0073] S5-1-1. Use drones to scan dispersed earth channels that have been damaged by water to determine the location of training water damage and its initial training terrain information;

[0074] S5-1-2. Combined with manual auxiliary marking, use discontinuity analysis technology and SURF algorithm to map the training water damage location, and extract the initial training water damage information and initial training channel information of the training water damage location;

[0075] S5-1-3, based on the initial training terrain information, the initial training water damage information, and the initial training channel information, the initial training operation information of the training water damage location is determined based on the geographical location;

[0076] S5-1-4, uniformly correct the position, time and unit of the initial training terrain information, the initial training water damage information, the initial training channel information and the initial training operation information to obtain the training water damage channel information, the training terrain information and the training operation information.

[0077] S5-2, constructing an initial decentralized earth channel water damage training database based on training water damage channel information, training terrain information and training operation information;

[0078] S5-3, performing data cleaning on all training data in the initial decentralized earth canal risk training database to obtain a decentralized earth canal water damage training database;

[0079] S5-4, using the training water damage location and training water damage information as labels of the water damage risk intelligent identification model;

[0080] S5-5. Input the training channel information, training terrain information, training operation information and labels in the decentralized earth channel water damage training database into the water damage risk intelligent identification model for training, adjust the number of hidden layers, the number of hidden units and the weight parameters of the water damage risk intelligent identification model, and obtain the trained water damage risk intelligent identification model.

[0081] S6. Input the processed channel information, terrain information and operation information into the water damage risk intelligent identification model to obtain the central location of the water damage prone area and its water damage information;

[0082] like Figure 2 As shown, the water damage information includes the water damage depth and water damage width of the water damage location.

[0083] The specific process of step S6 is:

[0084] The channel information, terrain information, and operation information of the dispersed earth channels to be evaluated are matched with the channel information, terrain information, and operation information of the dispersed earth channels that have suffered water damage through the intelligent identification model of water damage risk. The data consistent with the dispersed earth channels that have suffered water damage are regarded as areas prone to water damage, and the corresponding central location and its water damage information are output.

[0085] The water damage information includes the water damage depth and water damage width of the water damage location.

[0086] S7. Calculate the actual location of the water-prone area based on the spatial coordinates, flight attitude, and center location of the water-prone area of ​​the dispersed earth channel;

[0087] The calculation formula of step S7 is:

[0088] H=f×GSD / a

[0089] X=X0+H×tan(α)×cos(β)

[0090] Y=Y0+H×tan(α)×sin(β)

[0091] Among them, X and Y represent the actual position of the water-prone area, X0 and Y0 represent the spatial coordinates of the UAV, H represents the vertical distance from the UAV to the canal top at the center of the water-prone area, f represents the focal length, GSD represents the ground resolution, a represents the pixel size, α represents the angle between the center of the water-prone area and the UAV and the vertical, β represents the angle between the center of the water-prone area and the horizontal axis in the UAV spatial coordinates, tan(·), cos(·), and sin(·) represent the tangent function, cosine function, and sine function, respectively.

[0092] S8. Combining the processed operation information, terrain information, water damage information of water damage prone areas, and actual locations, conduct risk assessment on water damage prone areas to obtain corresponding risk assessment results.

[0093] Step S8 includes the following steps:

[0094] S8-1. Based on the depth and width of water damage in each water damage prone area, each water damage prone area is classified to obtain the corresponding water damage level; the water damage level is divided into low risk, medium risk, high risk and extremely dangerous;

[0095] If the water damage depth h is less than 0.5m and the water damage width w is less than 5m, the water damage level is low risk; if the water damage depth h∈[0.5m,1.5m) and the water damage width w∈[5m,10m), the water damage level is medium risk; if the water damage depth h∈[1.5m,3.0m) and the water damage width w∈[10m,20m), the water damage level is high risk; if the water damage depth h>3.0m and the water damage width w>20m, the water damage level is extremely dangerous.

[0096] S8-2. Based on the catchment area classification conditions, the dispersed soil channels to be evaluated are classified to obtain the corresponding catchment levels; the catchment levels are divided into low risk, medium risk, high risk and extremely dangerous;

[0097] According to the actual situation of dispersion and channel, the catchment area classification conditions are formulated. For example, if the catchment area is between 0 and 10 km 2 If the confluence area is between 10 and 50 km, the confluence level is low risk; if the confluence area is between 10 and 50 km 2 If the confluence area is between 50 and 100 km, the confluence level is medium risk. 2 If the confluence area exceeds 100km, the confluence level is high risk; if the confluence area exceeds 100km 2 , then the confluence level is extremely dangerous;

[0098] S8-3. Based on the classification conditions of precipitation intensity and rainfall duration, the dispersed soil channels to be evaluated are classified to obtain the corresponding rainfall levels; wherein the rainfall levels are divided into low risk, medium risk, high risk and extremely dangerous;

[0099] According to the actual situation of dispersion and channels and the precipitation intensity classification standards issued by the National Meteorological Administration, corresponding classification conditions are formulated. The rainfall intensity is divided into light rain, moderate rain, heavy rain, rainstorm and other levels, and the rainfall duration is divided into short-term rainfall (such as less than 1 hour), medium-term rainfall (such as 1 to 24 hours) and long-term rainfall (more than 24 hours). For example, if the rainfall intensity reaches the level of rainstorm and above and the rainfall intensity reaches the level of medium-term rainfall and above, the rainfall level is extremely dangerous; if the rainfall intensity is heavy rain and the rainfall intensity reaches the level of medium-term rainfall and above, the rainfall level is high risk; if the rainfall intensity is moderate rain and the rainfall intensity is medium-term rainfall, the rainfall level is medium risk; in all other cases, the rainfall level is low risk.

[0100] S8-4. Determine the risk assessment result based on the water damage level, runoff level and rainfall level. If the water damage level is extremely dangerous, the runoff level and rainfall level are both extremely dangerous, or the water damage level is high risk and either the runoff level or rainfall level is extremely dangerous, the risk assessment result is determined to be extremely dangerous, and corresponding preventive measures need to be formulated, etc.; if any two of the water damage level, runoff level and rainfall level are high risk, the risk assessment result is determined to be highly dangerous; if any two of the water damage level, runoff level and rainfall level are medium risk, the risk assessment result is determined to be medium risk; in all other cases, the risk assessment result is determined to be low risk.

[0101] In summary, the present invention proposes an intelligent assessment method for water damage risk of decentralized earth channels. Before the decentralized earth channels are damaged by water, the causal model and the deep learning network are used to accurately and quickly determine the locations in the decentralized earth channels that are prone to water damage, and the risk level is determined according to the classification standard. Therefore, protective measures can be reasonably designed in the engineering design to reduce or avoid the occurrence of water damage risks and effectively prevent water damage to the channels. In addition, during the on-site inspection during the operation period, key inspection locations can be set according to the identified water-prone locations, so as to discover the water damage locations early and avoid the increase of water damage due to omissions, thereby reducing the difficulty and cost of engineering repair. This has important engineering significance for the engineering design, safety assessment, hazard prevention and emergency rescue of decentralized earth channels.

Claims

1. An intelligent assessment method for water damage risk of decentralized earth channels, characterized by: The following steps are involved: S1. Use drones to inspect the dispersed soil channels to be evaluated and obtain the corresponding channel image data in real time; S2, extracting features from the channel image data to obtain channel information and terrain information; S3, obtaining the operation information of the dispersed soil channel to be evaluated; S4, performing data processing on the channel information, the terrain information, and the operation information to obtain processed channel information, terrain information, and operation information; S5. Construct an intelligent identification model for water damage risks; S6. Input the processed channel information, terrain information and operation information into the water damage risk intelligent identification model to obtain the central location of the water damage prone area and its water damage information; S7. Calculate the actual location of the water-prone area based on the spatial coordinates, flight attitude, and center location of the water-prone area of ​​the dispersed earth channel; S8. Combining the processed operation information, terrain information, water damage information of water damage prone areas, and actual locations, conduct risk assessment on water damage prone areas to obtain corresponding risk assessment results.

2. The method for intelligent assessment of water damage risk of dispersed earth channels according to claim 1 is characterized by: The step S1 comprises the following steps: S1-1. Use drones to inspect the dispersed soil channels to be evaluated, collect the corresponding initial channel image data and flight environment data, and upload them to the server in real time; S1-2. Based on the flight environment data, a data processing method is selected to perform image enhancement and geometric radiation correction on the initial channel image data to obtain processed channel image data, namely, channel image data.

3. The method for intelligent assessment of water damage risk of dispersed earth channels according to claim 1 is characterized by: The channel information includes channel height and slope ratio; the terrain information includes the catchment area and the initial height difference between the channel top and the outside of the channel; the operation information includes the channel operation time, precipitation intensity and rainfall duration; the water damage information includes the water damage depth and water damage width at the water damage location.

4. The method for intelligent assessment of water damage risk of dispersed earth channels according to claim 1 is characterized by: The data processing in step S4 includes data correction and data cleaning, wherein: The specific process of data correction is as follows: Performing position correction, time correction and unit correction on the channel information, terrain information and operation information to obtain corrected channel information, terrain information and operation information; The specific process of data cleaning is as follows: The corrected channel information, terrain information and operation information are screened, abnormal data are deleted, and missing data are completed using empirical values ​​or interpolation methods. The completed data and normal data are retained to obtain the processed channel information, terrain information and operation information.

5. The method for intelligent assessment of water damage risk of dispersed earth channels according to claim 1 is characterized by: The water damage risk intelligent identification model in step S5 adopts a deep learning network based on a causal model, including an input layer, multiple hidden layers and an output layer, and each hidden layer includes multiple hidden units.

6. The method for intelligent assessment of water damage risk of dispersed earth channels according to claim 5 is characterized by: The training process of the water damage risk intelligent identification model in step S5 includes the following steps: S5-1, obtaining the training water damage position, training water damage information, training channel information, training terrain information and training operation information of the dispersed earth channel that has suffered water damage; S5-2, constructing an initial decentralized earth channel water damage training database based on training water damage channel information, training terrain information and training operation information; S5-3, performing data cleaning on all training data in the initial decentralized earth canal risk training database to obtain a decentralized earth canal water damage training database; S5-4, using the training water damage location and training water damage information as labels of the water damage risk intelligent identification model; S5-5. Input the training channel information, training terrain information, training operation information and labels in the decentralized earth channel water damage training database into the water damage risk intelligent identification model for training, adjust the number of hidden layers, the number of hidden units and the weight parameters of the water damage risk intelligent identification model, and obtain the trained water damage risk intelligent identification model.

7. The method for intelligent assessment of water damage risk of dispersed earth channels according to claim 6 is characterized by: The step S5-1 comprises the following steps: S5-1-1. Use drones to scan dispersed earth channels that have been damaged by water to determine the location of training water damage and its initial training terrain information; S5-1-2. Combined with manual auxiliary marking, use discontinuity analysis technology and SURF algorithm to map the training water damage location, and extract the initial training water damage information and initial training channel information of the training water damage location; S5-1-3, based on the initial training terrain information, the initial training water damage information, and the initial training channel information, the initial training operation information of the training water damage location is determined based on the geographical location; S5-1-4, uniformly correct the position, time and unit of the initial training terrain information, the initial training water damage information, the initial training channel information and the initial training operation information to obtain the training water damage channel information, the training terrain information and the training operation information.

8. The method for intelligent assessment of water damage risk of dispersed earth channels according to claim 1 is characterized by: The specific process of step S6 is as follows: The channel information, terrain information, and operation information of the dispersed earth channels to be evaluated are matched with the channel information, terrain information, and operation information of the dispersed earth channels that have suffered water damage through the intelligent identification model of water damage risk. The data consistent with the dispersed earth channels that have suffered water damage are regarded as areas prone to water damage, and the corresponding central location and its water damage information are output.

9. The method for intelligent assessment of water damage risk of dispersed earth channels according to claim 1 is characterized by: The calculation formula of step S7 is: H=f×GSD / a X=X0+H×tan(α)×cos(β) Y=Y0+H×tan(α)×sin(β) Among them, X and Y represent the actual position of the water-prone area, X0 and Y0 represent the spatial coordinates of the UAV, H represents the vertical distance from the UAV to the canal top at the center of the water-prone area, f represents the focal length, GSD represents the ground resolution, a represents the pixel size, α represents the angle between the center of the water-prone area and the UAV and the vertical, β represents the angle between the center of the water-prone area and the horizontal axis in the UAV spatial coordinates, tan(·), cos(·), and sin(·) represent the tangent function, cosine function, and sine function, respectively.

10. The method for intelligent assessment of water damage risk of dispersed earth channels according to claim 8 is characterized by: The step S8 comprises the following steps: S8-1. Based on the depth and width of water damage in each water damage prone area, each water damage prone area is classified to obtain the corresponding water damage level; the water damage level is divided into low risk, medium risk, high risk and extremely dangerous; S8-2. Based on the catchment area classification conditions, the dispersed soil channels to be evaluated are classified to obtain the corresponding catchment levels; the catchment levels are divided into low risk, medium risk, high risk and extremely dangerous; S8-3. Based on the classification conditions of precipitation intensity and rainfall duration, the dispersed soil channels to be evaluated are classified to obtain the corresponding rainfall levels; wherein the rainfall levels are divided into low risk, medium risk, high risk and extremely dangerous; S8-4. Determine the risk assessment results based on the water damage level, runoff level and rainfall level.

Images