Water resource environment evaluation method and system based on machine vision

Abstract

The present application relates to the field of water resources environment monitoring, and more particularly to a water resources environment evaluation method and system based on machine vision, the steps of the method comprising: obtaining multi-modal water body visual data and dam structure feature data of a target reservoir; extracting floating trajectories of floating objects on the water surface, sediment-laden flow boundaries, and temperature difference boundaries based on the multi-modal water body visual data; extracting water environment disturbance characteristics of sediment-laden flow on different dam body regions; determining the risk amplification degree of dam structure sensitivity to water environment disturbance; analyzing dam structure risks after amplification of water environment disturbance by dam structure sensitivity according to the water environment disturbance intensity of each dam body region and the risk amplification degree of dam structure sensitivity to water environment disturbance, and generating a water resources environment evaluation result of the target reservoir. The present application quantifies water environment disturbance risks using multi-modal water body visual features, thereby improving the accuracy of reservoir dam structure risk identification and early warning.

Description

Technical Field

[0001] This invention relates to the field of water resources and environment monitoring, and in particular to a water resources and environment assessment method and system based on machine vision. Background Technology

[0002] As a crucial infrastructure for water resource regulation and flood control, the water environment of reservoirs not only affects water quality and the safety of water resource utilization, but also influences the operational status of the dam structure through water erosion and sediment transport. Especially under conditions of heavy rainfall, a sudden increase in upstream sediment load, and rapid changes in reservoir water levels, abnormal water flows can easily form in the reservoir, exhibiting turbidity, sediment content, and temperature distinct from the surrounding water bodies. As these abnormal flows propagate towards the vicinity of the dam, they can cause localized sediment deposition, erosion, changes in water pressure, or alterations in seepage conditions. Therefore, timely assessment of reservoir water environment anomalies and their potential impacts on the dam area is necessary.

[0003] Existing water resource environmental assessment methods typically obtain water body state parameters through water quality monitoring stations, turbidity sensors, temperature sensors, flow meters, or manual sampling. While these methods can obtain water body parameters near the monitoring points, they are limited by the location and number of monitoring devices, making it difficult to continuously reflect the spatial distribution and migration process of abnormal water flows. When sediment-laden water flows exhibit significant spatial heterogeneity, local monitoring data cannot accurately represent the boundary changes, movement direction, and diffusion differences between different areas.

[0004] With the development of machine vision technology, imaging techniques such as visible light, thermal infrared, and hyperspectral imaging are gradually being applied to the field of water environment monitoring. Visible light images can intuitively track the movement trajectory of floating objects on the water surface, thermal infrared images can reflect the temperature difference distribution of large water bodies, and hyperspectral images can perform detailed inversion of water turbidity and sediment concentration. However, existing visual assessment methods mostly rely on single-modal data, such as using visible light images alone to extrapolate surface flow fields, or using thermal infrared images alone to identify the boundaries between hot and cold water masses. A single data source cannot simultaneously acquire the relationship between sediment-laden water flow boundaries, temperature difference boundaries, and the movement of surface floating objects. Especially during the process of sediment-laden water flowing downwards from the surface, its surface optical characteristics gradually disappear. Although the temperature difference caused by underwater movement can be captured by thermal infrared, it is difficult to quantify the directional water environment disturbance caused by sediment-laden water flow to different dam areas based on single or simply superimposed multi-source images, resulting in incomplete assessment information.

[0005] Therefore, the existing technology has at least the following shortcomings: (1) It is difficult to accurately identify the state changes of sediment-laden water flow from the surface to the underwater submerged using multimodal visual data; (2) It is difficult to quantify the intensity and direction of water environment disturbance of sediment-laden water flow to different dam areas based on the relative motion relationship between the floating trajectory, the boundary of sediment-laden water flow and the temperature difference boundary; (3) There is a lack of technical means to correlate the characteristics of water environment disturbance with the sensitive direction of dam structure, which makes it difficult for the water resource environment assessment results to reflect the dam structure risk formed after the water environment anomaly is amplified by the sensitivity of dam structure. Summary of the Invention

[0006] To overcome the defects and shortcomings of existing technologies, this invention provides a water resource environment assessment method and system based on machine vision. It utilizes multimodal water body visual features to quantify the risk of water environment disturbance, thereby improving the accuracy of reservoir dam structure risk identification and early warning.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a water resource environment assessment method based on machine vision, comprising the following steps: S1. Acquire multimodal visual data of the target reservoir and structural feature data of the dam. The multimodal visual data of the water body includes visible light image data, hyperspectral image data and thermal infrared image data. S2. Extract the floating trajectory of floating objects on the water surface based on visible light image data, analyze the turbidity distribution of water body and extract the boundary of sediment-laden water flow based on hyperspectral image data, and analyze the temperature difference distribution of water body and extract the temperature difference boundary based on thermal infrared image data. S3. Based on the migration trend of the sediment-laden flow boundary, determine the submerged position where the sediment-laden flow changes from the surface flow state to the underwater submerged state. Combined with the floating trajectory, the degree of motion separation between the turbid water boundary and the temperature difference boundary, extract the water environment disturbance characteristics of the sediment-laden flow to different dam areas. The water environment disturbance characteristics include the water environment disturbance intensity and the water environment disturbance direction. S4. Based on the structural characteristic data of the dam body, analyze the structurally sensitive direction corresponding to each dam body area, and determine the degree of risk amplification of water environment disturbance by the structural sensitivity of the dam body based on the matching relationship between the structurally sensitive direction and the direction of water environment disturbance. S5. Based on the intensity of water environment disturbance in each dam area and the degree of risk amplification of water environment disturbance by the dam structure sensitivity, analyze the dam structure risk after the water environment disturbance is amplified by the dam structure sensitivity, and generate the water resource environment assessment results of the target reservoir.

[0008] Furthermore, in step S3, the intrusion location where the sediment-laden water flow transitions from a surface-following state to an underwater intrusion state is determined based on the migration trend of the sediment-laden water flow boundary. Specific steps include: S31. Divide the boundary of the sediment-laden flow into boundary units, and determine the migration velocity, migration direction and boundary curvature change rate of each boundary unit based on the positional changes of each boundary unit at continuous time intervals. S32. Extract the surface water reference velocity and the main direction of the water surface corresponding to the location of each boundary unit based on the floating trajectory. Combine the migration velocity and migration direction of each boundary unit to determine the degree of deviation of the migration velocity and the degree of deviation of the migration direction of each boundary unit. S33. By weighted summing of the degree of deviation of migration speed, the degree of deviation of migration direction and the rate of change of boundary curvature for each boundary element, the degree of boundary migration anomaly for each boundary element is obtained. S34. Boundary units with a boundary migration anomaly degree greater than the preset boundary migration anomaly threshold are taken as target boundary units. The spatial locations corresponding to the target boundary units are clustered, and the center locations of each cluster are taken as the intrusion locations where the sediment-laden water flow changes from the surface flow state to the underwater intrusion state.

[0009] Further, the specific steps of step S32 include: projecting the migration velocity of each boundary unit onto the corresponding mainstream water direction to obtain the forward migration velocity of each boundary unit, and taking the ratio of the absolute value of the difference between the forward migration velocity and the corresponding surface water reference velocity to the surface water reference velocity as the degree of deviation of the migration velocity corresponding to each boundary unit; calculating the deviation angle between the migration direction of each boundary unit and the corresponding mainstream water direction, and taking the ratio of the deviation angle to the average value of all deviation angles as the degree of deviation of the migration direction corresponding to each boundary unit.

[0010] Furthermore, step S3 extracts the water environment disturbance characteristics of sediment-laden water flow on different dam areas, specifically including: S35. By performing time synchronization and spatial registration on visible light image data, hyperspectral image data and thermal infrared image data at consecutive times, the floating trajectory, sediment-laden water flow boundary and temperature difference boundary are mapped to the same spatial coordinate system of the target reservoir. S36. Construct time series analysis windows centered on each intrusion location, extract the spatial positions of the floating trajectory, sediment-laden water flow boundary, and temperature difference boundary at continuous time intervals within each time series analysis window, and use spline curve fitting to obtain the floating trajectory, sediment-laden water flow boundary migration trajectory, and temperature difference boundary migration trajectory, and determine the first motion separation degree between the floating trajectory and the sediment-laden water flow boundary, the second motion separation degree between the floating trajectory and the temperature difference boundary, and the cooperative migration degree between the sediment-laden water flow boundary and the temperature difference boundary, respectively. Take the direction of the sediment-laden water flow boundary migration trajectory as the direction of water environment disturbance. S37. By weighted summing of the first motion separation degree, the second motion separation degree, and the cooperative migration degree corresponding to each intrusion location, the water environment disturbance characteristic value corresponding to each time moment is obtained, and the average value of the water environment disturbance characteristic value within the corresponding time series analysis window is taken as the water environment disturbance intensity of the sediment-laden water flow moving from the surface to the underwater submersion at each intrusion location.

[0011] Furthermore, the specific steps of step S36 include: S361. Based on the spatial coordinates of the floating trajectory, the sediment-laden flow boundary migration trajectory, and the temperature difference boundary migration trajectory at the same time, calculate the first relative position vector of the sediment-laden flow boundary migration trajectory relative to the floating trajectory, the second relative position vector of the temperature difference boundary migration trajectory relative to the floating trajectory, and the third relative position vector of the temperature difference boundary migration trajectory relative to the sediment-laden flow boundary migration trajectory. S362. Subtract the first relative position vector and the second relative position vector at each time point from the initial relative position vector corresponding to the start time of the time series analysis window to obtain the first relative separation displacement of the sediment-laden water flow boundary relative to the floating trajectory and the second relative separation displacement of the temperature difference boundary relative to the floating trajectory. S363. Calculate the root mean square values ​​of the first relative separation displacement modulus and the second relative separation displacement modulus within the time series analysis window, and calculate the ratio with the cumulative migration length of the floating trajectory within the time series analysis window to obtain the first degree of motion separation between the floating trajectory and the boundary of the sediment-laden water flow and the second degree of motion separation between the floating trajectory and the temperature difference boundary. S364. Subtract the third relative position vector at each time point from the initial third relative position vector corresponding to the start time of the time series analysis window to obtain the relative offset displacement of the temperature difference boundary relative to the sediment-laden water flow boundary, and calculate the root mean square value of the relative offset displacement magnitude within the time series analysis window. S365. Calculate the ratio of the root mean square value to the average cumulative migration length of the sediment-laden flow boundary migration trajectory and the temperature difference boundary migration trajectory to obtain the relative offset between the sediment-laden flow boundary and the temperature difference boundary. Take the reciprocal of the sum of the value 1 and the relative offset as the degree of cooperative migration between the sediment-laden flow boundary and the temperature difference boundary.

[0012] Furthermore, in step S4, determining the degree to which the dam structure's sensitivity amplifies the risk of water environment disturbance includes the following specific steps: S41. Extract the crack extension direction and structural joint extension direction in each dam area based on the dam structure feature data, and use the normal direction of the crack extension direction and structural joint extension direction as the local structural sensitive direction of the corresponding dam area. S42. The absolute value of the cosine of the angle between the sensitive direction of the local structure and the direction of water environment disturbance is taken as the direction matching degree of each dam area. S43. Determine the structural sensitivity of each dam area based on the distribution range of cracks and structural joints within the corresponding dam area. S44. By weighting and summing the structural sensitivity through directional matching degree, we can obtain the degree of risk amplification of the dam structure sensitivity to the corresponding water environment disturbance.

[0013] Furthermore, the specific steps for generating the water resource environment assessment results of the target reservoir in step S5 include: S51. The sum of the value 1 and the risk amplification degree of the dam structure sensitivity to water environment disturbance is taken as the risk amplification factor, and the product of the water environment disturbance intensity and the risk amplification factor is taken as the dam structure risk corresponding to each dam area. S52. Output the water environment disturbance characteristics of sediment-laden water flow to different dam areas as the water resource environment assessment results of the target reservoir. When the dam structure risk is greater than the preset dam structure risk threshold, a dam structure risk warning is issued. When the dam structure risk is less than or equal to the preset dam structure risk threshold, no dam structure risk warning is issued.

[0014] Secondly, the present invention provides a machine vision-based water resource environment assessment system, comprising: The data acquisition module is used to acquire multimodal visual data of the target reservoir and structural feature data of the dam. The multimodal visual data of the water body includes visible light image data, hyperspectral image data and thermal infrared image data. The boundary extraction module is used to extract the floating trajectory of floating objects on the water surface based on visible light image data, analyze the turbidity distribution of water body based on hyperspectral image data and extract the boundary of sediment-laden water flow, and analyze the temperature difference distribution of water body based on thermal infrared image data and extract the temperature difference boundary. The water environment disturbance extraction module is used to determine the intrusion location of the sediment-laden water flow from the surface drift state to the underwater intrusion state based on the migration trend of the sediment-laden water flow boundary. It also extracts the water environment disturbance characteristics of the sediment-laden water flow on different dam areas by combining the floating trajectory, the degree of motion separation between the turbid water boundary and the temperature difference boundary. The water environment disturbance characteristics include the water environment disturbance intensity and the water environment disturbance direction. The dam structure sensitivity analysis module is used to analyze the structural sensitivity direction corresponding to each dam area based on the dam structure characteristic data, and to determine the degree of risk amplification of water environment disturbance by the dam structure sensitivity based on the matching relationship between the structural sensitivity direction and the water environment disturbance direction. The assessment results output module is used to analyze the risk of water environment disturbance amplified by the dam structure sensitivity after the water environment disturbance is amplified, based on the intensity of water environment disturbance in each dam area and the degree of risk amplification of water environment disturbance by the dam structure sensitivity, and to generate the water resource environment assessment results of the target reservoir.

[0015] Thirdly, the present invention provides an electronic device, comprising: a processor and a memory, wherein the memory stores a computer program that can be called by the processor, and the processor executes a machine vision-based water resource environment assessment method by calling the computer program stored in the memory.

[0016] Fourthly, the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform a machine vision-based water resources environment assessment method.

[0017] Compared with the prior art, the present invention has the following advantages and beneficial effects: This invention integrates visible light, hyperspectral, and thermal infrared images to identify the state changes of sediment-laden water flow from the surface to the submerged depths. Based on the motion separation relationship between the floating trajectory, the sediment-laden water flow boundary, and the temperature difference boundary, it quantifies the intensity and direction of water environment disturbance in different dam areas. Simultaneously, by combining the structurally sensitive directions corresponding to dam cracks and structural joints, it determines the degree to which the dam structure sensitivity amplifies the risk of water environment disturbance. This enables a correlation assessment of water anomalies, water environment disturbances, and dam structure risks, improving the accuracy of reservoir dam structure risk identification and early warning. Attached Figure Description

[0018] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a flowchart illustrating the machine vision-based water resource environment assessment method provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the process for extracting water environment disturbance characteristics provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of the machine vision-based water resources environment assessment system provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0019] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0020] Please see Figure 1 , Figure 1 This is a schematic diagram of the overall process of the machine vision-based water resource environment assessment method provided in this embodiment of the invention, which specifically includes the following steps: S1. Acquire multimodal visual data of the target reservoir and structural feature data of the dam. The multimodal visual data of the water body includes visible light image data, hyperspectral image data and thermal infrared image data. Specifically, visible light cameras, hyperspectral imagers, and thermal infrared cameras with overlapping fields of view are deployed on the bank of the target reservoir, the top of the dam, or on a drone-mounted platform. Continuous imaging of the reservoir surface and the water area in front of the dam is carried out according to a unified sampling cycle. Through time stamp synchronization, camera calibration, and geographic coordinate registration, visible light image data reflecting changes in the position of floating objects on the water surface, hyperspectral image data reflecting the spectral reflectance characteristics and turbidity distribution of the water body, and thermal infrared image data reflecting the temperature differences of the water surface are obtained, forming multimodal visual data of the water body. Dam structural feature data can be obtained by retrieving dam design drawings and combining them with 3D scanning data collected by drones or dam inspection equipment. The spatial location, dam orientation, structural joint location and extension direction, crack location and extension direction, and corresponding distribution range of each dam area are extracted.

[0021] S2. Extract the floating trajectory of floating objects on the water surface based on visible light image data, analyze the turbidity distribution of water body and extract the boundary of sediment-laden water flow based on hyperspectral image data, and analyze the temperature difference distribution of water body and extract the temperature difference boundary based on thermal infrared image data. Specifically, by performing distortion correction and water surface region segmentation on continuous visible light images, floating objects are identified based on the differences in color and texture features between the floating objects and the background water. The center coordinates, contours, and appearance features of the floating objects are extracted, and a target correspondence is established based on the spatial distance and appearance similarity between floating objects at adjacent time points. The center coordinates of the same floating object at consecutive time points are connected in chronological order to obtain the corresponding floating trajectory. By performing radiometric and atmospheric correction on hyperspectral images, the spectral reflectance of each water surface pixel is extracted. Turbidity characterization values ​​are constructed based on characteristic bands sensitive to suspended sediment, and the turbidity of on-site water samples is used to further characterize the turbidity. The turbidity detection results establish a mapping relationship between spectral characterization values ​​and water turbidity, obtaining the water turbidity distribution. Continuous areas with turbidity greater than the surrounding background water are identified as sediment-laden flow areas, and their outer contours are extracted and smoothed to obtain the sediment-laden flow boundary. By performing non-uniformity correction and temperature inversion on thermal infrared images, the water surface temperature distribution is obtained. Using the local background water temperature or the average water temperature of the monitoring area as the reference water temperature, the temperature difference between each water surface position and the reference water temperature is calculated. Areas with continuous temperature differences that are significantly different from the surrounding water are identified as temperature difference anomalous areas, and the outer contour of these temperature difference anomalous areas is extracted as the temperature difference boundary.

[0022] S3. Based on the migration trend of the sediment-laden flow boundary, determine the submerged position where the sediment-laden flow changes from the surface flow state to the underwater submerged state. Combined with the floating trajectory, the degree of motion separation between the turbid water boundary and the temperature difference boundary, extract the water environment disturbance characteristics of the sediment-laden flow to different dam areas. The water environment disturbance characteristics include the water environment disturbance intensity and the water environment disturbance direction. When sediment-laden water flows at the surface, their surface projection boundary typically maintains a similar migration speed and direction to the surface water flow represented by floating objects, and the boundary morphology is relatively stable. However, when the sediment-laden water flow gradually submerges due to sediment concentration differences, temperature differences, or water density stratification, the projection of the sediment-laden water flow boundary on the water surface exhibits changes such as reduced forward migration speed, migration direction deviating from the mainstream surface flow direction, and increased local boundary contraction or bending. This embodiment uses the sediment-laden water flow boundary extracted from hyperspectral images as a spatial representation of abnormal sediment-laden water bodies, and uses the surface water flow state reflected by the movement of floating objects in visible light images as a reference. By comparing the relative motion differences between the sediment-laden water flow boundary and the surface water flow, the location where the sediment-laden water flow transitions from surface transport to submerged migration is identified. Furthermore, by combining the temperature difference boundary extracted from thermal infrared images, it can be verified whether sediment anomalies and temperature anomalies jointly detach from the surface water flow near corresponding locations, thereby reducing the impact of wind, surface fluctuations, local turbidity diffusion, or other factors. Misjudgments caused by image noise; by determining the intrusion location, a clear spatial benchmark is provided for subsequent construction of time-series analysis windows, analysis of multimodal boundary motion separation relationships, and extraction of water environment disturbance intensity and direction. This avoids misjudging the intrusion of sediment-laden water flow as the dissipation of sediment-laden water flow simply based on the shrinkage of the sediment-laden area on the water surface, thus improving the accuracy of underwater migration process identification and water environment risk assessment. In step S3, the intrusion location where the sediment-laden water flow transitions from a surface drift state to an underwater intrusion state is determined based on the migration trend of the sediment-laden water flow boundary. The specific steps include: S31. Divide the sediment-laden flow boundary into boundary units, and determine the migration velocity, migration direction, and boundary curvature change rate of each boundary unit based on its positional changes at consecutive time points. Specifically, perform radiometric correction, geometric correction, and spatial registration on the hyperspectral images at consecutive time points. Segment the sediment-laden flow region at each time point based on the difference between the turbidity of the water body and the background water body, and extract the outer contour of the sediment-laden flow region as the sediment-laden flow boundary. Set boundary feature points on the sediment-laden flow boundary at equal arc length intervals, and form a boundary unit with adjacent boundary feature points. For any boundary unit, calculate the difference in its center position between adjacent time points, and... The ratio of the difference in center position to the time interval is used as the migration velocity vector of the boundary unit. The magnitude of the migration velocity vector is used as the migration velocity, and the azimuth of the migration velocity vector in the unified spatial coordinate system is used as the migration direction. Local spline curve fitting is performed on the boundary feature points within the boundary unit. The curvature of each boundary feature point is calculated based on the first and second derivatives of the fitted curve. The average value of the absolute curvature of each boundary feature point within the boundary unit is used as the average curvature of the boundary unit at the corresponding time. Furthermore, the ratio of the absolute value of the difference in average curvature between adjacent time points to the time interval is used as the rate of change of the boundary curvature of the boundary unit, so as to characterize the speed of local bending, contraction or abrupt morphological changes of the sediment-laden water flow boundary. S32. Extract the surface water flow reference velocity and the main water flow direction corresponding to the location of each boundary unit based on the floating trajectory. Combine the migration velocity and migration direction of each boundary unit to determine the degree of migration velocity deviation and migration direction deviation of each boundary unit. The specific steps include: projecting the migration velocity of each boundary unit onto the corresponding main water flow direction to obtain the forward migration velocity of each boundary unit, and taking the ratio of the absolute value of the difference between the forward migration velocity and the corresponding surface water flow reference velocity to the surface water flow reference velocity as the degree of migration velocity deviation of each boundary unit; calculating the deviation angle between the migration direction of each boundary unit and the corresponding main water flow direction, and taking the ratio of the deviation angle to the average value of all deviation angles as the degree of migration direction deviation of each boundary unit. S33. By weighted summing of the migration velocity deviation, migration direction deviation and boundary curvature change rate corresponding to each boundary unit, the boundary migration anomaly degree corresponding to each boundary unit is obtained. The greater the boundary migration anomaly degree, the more likely the boundary unit has the characteristics of deceleration or acceleration relative to the surface water flow, deflection of the movement direction and rapid change of the boundary morphology. It is more likely to correspond to a local area where the sediment-laden water flow changes from the surface flow state to the underwater submerged state. S34. Boundary units with anomalies exceeding a preset boundary migration anomaly threshold are designated as target boundary units. The spatial locations of these target boundary units are clustered, and the center of each cluster is used as the entry point for the sediment-laden water flow to transition from a surface-following state to a submerged state. The preset boundary migration anomaly threshold can be determined based on the sum of the mean and standard deviation of the boundary migration anomaly degree under historical stable surface-following states. The clustering process can be as follows: extract the center coordinates of each target boundary unit and calculate the spatial distance between any two target boundary units, treating each target boundary unit as an independent cluster. The two closest clusters are merged sequentially according to the minimum spatial distance between them, and the inter-cluster distance corresponding to each merge is recorded. The growth rate between two adjacent merge distances is calculated, and the position with the largest growth rate is taken as the cluster boundary point. The clusters of target boundary units formed before this cluster boundary point are taken as the clustering result, and isolated units that have not formed stable clusters with other target boundary units are removed. The degree of boundary migration anomaly of each target boundary unit is used as the position weight, and the center coordinates of each target boundary unit in the same cluster are weighted and averaged to obtain the corresponding cluster center. The cluster center is determined as the intrusion position of the sediment-laden water flow.

[0023] Please see Figure 2 , Figure 2 This is a schematic flowchart of the process for extracting water environment disturbance characteristics provided in an embodiment of the present invention. Step S3 extracts the water environment disturbance characteristics of sediment-laden water flow on different dam areas. The specific steps include: S35. By performing time synchronization and spatial registration on visible light image data, hyperspectral image data and thermal infrared image data at consecutive times, the floating trajectory, sediment-laden water flow boundary and temperature difference boundary are mapped to the same spatial coordinate system of the target reservoir. S36. Construct time series analysis windows centered on each intrusion location, extract the spatial positions of the floating trajectory, sediment-laden water flow boundary, and temperature difference boundary at continuous time intervals within each time series analysis window, and use spline curve fitting to obtain the floating trajectory, sediment-laden water flow boundary migration trajectory, and temperature difference boundary migration trajectory, and determine the first motion separation degree between the floating trajectory and the sediment-laden water flow boundary, the second motion separation degree between the floating trajectory and the temperature difference boundary, and the cooperative migration degree between the sediment-laden water flow boundary and the temperature difference boundary, respectively. Take the direction of the sediment-laden water flow boundary migration trajectory as the direction of water environment disturbance. S37. By weighted summing of the first motion separation degree, the second motion separation degree, and the cooperative migration degree corresponding to each intrusion location, the water environment disturbance characteristic value corresponding to each time moment is obtained, and the average value of the water environment disturbance characteristic value within the corresponding time series analysis window is taken as the water environment disturbance intensity of the sediment-laden water flow moving from the surface to the underwater submersion at each intrusion location.

[0024] The specific steps of step S36 include: S361. Based on the spatial coordinates of the floating trajectory, the sediment-laden flow boundary migration trajectory, and the temperature difference boundary migration trajectory at the same time, calculate the first relative position vector of the sediment-laden flow boundary migration trajectory relative to the floating trajectory, the second relative position vector of the temperature difference boundary migration trajectory relative to the floating trajectory, and the third relative position vector of the temperature difference boundary migration trajectory relative to the sediment-laden flow boundary migration trajectory. S362. Subtract the first relative position vector and the second relative position vector at each time point from the initial relative position vector corresponding to the start time of the time series analysis window to obtain the first relative separation displacement of the sediment-laden water flow boundary relative to the floating trajectory and the second relative separation displacement of the temperature difference boundary relative to the floating trajectory. S363. Calculate the root mean square values ​​of the first relative separation displacement modulus and the second relative separation displacement modulus within the time series analysis window, and calculate the ratio with the cumulative migration length of the floating trajectory within the time series analysis window to obtain the first degree of motion separation between the floating trajectory and the boundary of the sediment-laden water flow and the second degree of motion separation between the floating trajectory and the temperature difference boundary. S364. Subtract the third relative position vector at each time point from the initial third relative position vector corresponding to the start time of the time series analysis window to obtain the relative offset displacement of the temperature difference boundary relative to the sediment-laden water flow boundary, and calculate the root mean square value of the relative offset displacement magnitude within the time series analysis window. S365. Calculate the ratio of the root mean square value to the average cumulative migration length of the sediment-laden flow boundary migration trajectory and the temperature difference boundary migration trajectory to obtain the relative offset between the sediment-laden flow boundary and the temperature difference boundary. Take the reciprocal of the sum of the value 1 and the relative offset as the degree of cooperative migration between the sediment-laden flow boundary and the temperature difference boundary.

[0025] S4. Based on the structural characteristic data of the dam body, analyze the structurally sensitive direction corresponding to each dam body area, and determine the degree of risk amplification of water environment disturbance by the structural sensitivity of the dam body based on the matching relationship between the structurally sensitive direction and the direction of water environment disturbance. Dam cracks and structural joints exhibit a clear directionality. For dam areas where crack opening, joint opening and closing, and seepage channel expansion are the main risks, when the direction of water environment disturbance is close to the normal direction of the crack or structural joint, the component of the disturbance in the normal direction of the weak structure is larger, making it easier to cause increased crack opening, enhanced joint permeability, or local water pressure transmission. When the direction of water environment disturbance is close to the extension direction of the crack or structural joint, the component acting on its normal direction is relatively smaller, and the structural risk amplification effect is relatively weaker. Based on the spatial distribution of dam cracks and structural joints, the local structurally sensitive direction is determined, and the structural sensitivity is corrected by the matching relationship between the structurally sensitive direction and the direction of water environment disturbance. This ensures that the same water environment disturbance intensity yields different risk amplification results when acting on different dam areas or in different directions, improving the pertinence and accuracy of the analysis of water environment disturbance risk transmission to the dam structure. Step S4 determines the degree of risk amplification of water environment disturbance by the dam structure sensitivity. The specific steps include: S41. Based on the dam structure feature data, extract the crack extension direction and structural joint extension direction within each dam area. Use the normals of the crack extension direction and structural joint extension direction as the local structural sensitive directions of the corresponding dam area. Specifically, divide the dam into multiple dam areas based on the dam structure feature data, and extract the contour point set corresponding to the cracks and structural joints within each dam area. Perform skeletonization processing on the contour point set of each crack and structural joint to obtain the corresponding centerline. Use principal component analysis or linear fitting to determine the crack extension direction and structural joint extension direction based on the coordinate distribution of each point on the centerline. In the dam surface tangent plane of the corresponding dam area, construct unit normal vectors perpendicular to the crack extension direction and structural joint extension direction, respectively, and use the unit normal vectors as the local structural sensitive directions of the corresponding cracks or structural joints. The local structural sensitive direction only represents the sensitive axis and does not distinguish between the positive and negative directions of the normal vector. S42. The absolute value of the cosine of the angle between the sensitive direction of the local structure and the direction of water environment disturbance is taken as the direction matching degree corresponding to each dam area. The value of the direction matching degree ranges from 0 to 1. The closer the value is to 1, the closer the water environment disturbance is to the normal action along the crack or structural joint. The closer the value is to 0, the closer the water environment disturbance is to the extension direction along the crack or structural joint. S43. Based on the distribution range of cracks and structural joints within the corresponding dam area, determine the structural sensitivity of each dam area. Specifically, calculate the actual distribution area of ​​each crack and structural joint within the corresponding dam area. The ratio of the distribution area of ​​each crack or structural joint to the area of ​​the corresponding dam area is taken as the local structural sensitivity of the crack or structural joint. The greater the local structural sensitivity, the greater the distribution range of the corresponding weak structure within the dam area. S44. By weighting and summing the structural sensitivity through directional matching degree, we can obtain the degree of risk amplification of the dam structure sensitivity to the corresponding water environment disturbance.

[0026] S5. Based on the intensity of water environment disturbance in each dam area and the degree of risk amplification of water environment disturbance by the dam structure sensitivity, analyze the dam structure risk after the water environment disturbance is amplified by the dam structure sensitivity, and generate the water resource environment assessment results of the target reservoir. By correlating and outputting the intensity and direction of water environment disturbance caused by sediment-laden water flow to different dam areas, and the resulting dam structural risk amplified by the dam structure's sensitivity, the water resource environment assessment results not only reflect the water anomaly itself but also characterize the regional differences in the transmission of water environment disturbance to dam structural risk. By using a risk amplification factor to correct the intensity of water environment disturbance, a uniform risk assessment is avoided for the same disturbance intensity. The risk amplification effect is highlighted when weak structures such as cracks and structural joints match the disturbance direction, thereby accurately identifying dam areas requiring key attention. This achieves a unified assessment of water environment status and dam structural risk early warning, providing a basis for regional inspections, operation scheduling, and safety management of the target reservoir. Step S5 generates the water resource environment assessment results for the target reservoir, and the specific steps include: S51. The sum of the value 1 and the risk amplification degree of the dam structure sensitivity to water environment disturbance is taken as the risk amplification factor, and the product of the water environment disturbance intensity and the risk amplification factor is taken as the dam structure risk corresponding to each dam area. S52. Output the water environment disturbance characteristics of sediment-laden water flow to different dam areas as the water resource environment assessment results of the target reservoir. When the dam structure risk is greater than the preset dam structure risk threshold, a dam structure risk warning is issued. When the dam structure risk is less than or equal to the preset dam structure risk threshold, no dam structure risk warning is issued.

[0027] In this embodiment of the invention, the determination of parameters such as weighted weights, preset boundary migration anomaly thresholds, and preset dam structure risk thresholds can be achieved by: constructing a dataset by acquiring multimodal water body visual data of the target reservoir and dam structure feature data, substituting these into the dataset to calculate the degree of boundary migration anomaly and dam structure risk, and simultaneously acquiring expert judgments on the degree of boundary migration anomaly and dam structure risk. The calculated degree of boundary migration anomaly, dam structure risk, and judgment results are then imported into fitting software, and the weighted weights, preset boundary migration anomaly thresholds, and preset dam structure risk thresholds that meet the maximum judgment accuracy are output.

[0028] Please see Figure 3 , Figure 3 This is a schematic diagram of the structure of a machine vision-based water resource environment assessment system provided in an embodiment of the present invention, including: The data acquisition module 210 is used to acquire multimodal visual data of the target reservoir and structural feature data of the dam. The multimodal visual data of the water body includes visible light image data, hyperspectral image data and thermal infrared image data. The boundary extraction module 220 is used to extract the floating trajectory of floating objects on the water surface based on visible light image data, analyze the turbidity distribution of water body based on hyperspectral image data and extract the boundary of sediment-laden water flow, and analyze the temperature difference distribution of water body based on thermal infrared image data and extract the temperature difference boundary. The water environment disturbance extraction module 230 is used to determine the submersion position of the sediment-laden water flow from the surface flowing state to the underwater submersion state based on the migration trend of the sediment-laden water flow boundary. It also extracts the water environment disturbance characteristics of the sediment-laden water flow to different dam areas by combining the floating trajectory, the degree of motion separation between the turbid water boundary and the temperature difference boundary. The water environment disturbance characteristics include the water environment disturbance intensity and the water environment disturbance direction. The dam structure sensitivity analysis module 240 is used to analyze the structural sensitivity direction corresponding to each dam area based on the dam structure characteristic data, and determine the degree of risk amplification of water environment disturbance by the dam structure sensitivity according to the matching relationship between the structural sensitivity direction and the water environment disturbance direction. The assessment result output module 250 is used to analyze the risk of water environment disturbance amplified by the dam structure sensitivity after the water environment disturbance is amplified based on the intensity of water environment disturbance in each dam area and the degree of risk amplification of water environment disturbance by the dam structure sensitivity, and generate the water resource environment assessment results of the target reservoir.

[0029] The steps for implementing the corresponding functions of each parameter and each unit module in the machine vision-based water resources and environment assessment system of the present invention described above can be referred to the parameters and steps in the embodiments of the machine vision-based water resources and environment assessment method described above, and will not be repeated here.

[0030] Please refer to Figure 4The present invention also provides an electronic device 300, including a memory 310, a processor 320, and a communication bus 330; the memory 310 and the processor 320 are connected via the communication bus 330. The memory 310 stores a water resource environment assessment method based on machine vision as provided in the above embodiments, which can be loaded and executed by the processor 320.

[0031] The memory 310 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 310 may include a program storage area and a data storage area. The program storage area may store instructions for implementing an operating system, instructions for at least one function, and instructions for implementing the machine vision-based water resource environment assessment method provided in the above embodiments, etc. The data storage area may store data involved in the machine vision-based water resource environment assessment method provided in the above embodiments, etc.

[0032] Processor 320 may include one or more processing cores. Processor 320 executes instructions, programs, code sets, or instruction sets stored in memory 310, and calls data stored in memory 310 to perform various functions and process data according to the present invention. Processor 320 may be at least one of the following: Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), Controller, Microcontroller, and Microprocessor. It is understood that, for different devices, the electronic devices used to implement the functions of processor 320 may also be other types, and the embodiments of the present invention do not specifically limit this.

[0033] The communication bus 330 may include a path for transmitting information between the aforementioned components. The communication bus 330 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The communication bus 330 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4The symbol is represented by a single double arrow, but this does not mean that there is only one bus or one type of bus.

[0034] This invention provides a computer-readable storage medium storing a computer program that can be loaded by a processor and executed as described in the above embodiments, representing a machine vision-based water resources environment assessment method.

[0035] In this embodiment of the invention, the computer-readable storage medium can be a tangible device that holds and stores instructions used by an instruction execution device. The computer-readable storage medium can be, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any combination thereof. Specifically, the computer-readable storage medium can be a portable computer disk, a hard disk, a USB flash drive, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), lectern random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory stick, floppy disk, optical disk, magnetic disk, mechanical encoding device, or any combination thereof.

[0036] 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 process, method, article, or apparatus.

[0037] The above description is merely a preferred embodiment of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this application is not limited to the technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions claimed in this invention.

Claims

1. A machine vision-based water resource environment assessment method, characterized in that, Includes the following steps: S1. Acquire multimodal visual data of the target reservoir and structural feature data of the dam. The multimodal visual data of the water body includes visible light image data, hyperspectral image data and thermal infrared image data. S2. Extract the floating trajectory of floating objects on the water surface based on visible light image data, analyze the turbidity distribution of water body and extract the boundary of sediment-laden water flow based on hyperspectral image data, and analyze the temperature difference distribution of water body and extract the temperature difference boundary based on thermal infrared image data. S3. Based on the migration trend of the sediment-laden flow boundary, determine the submerged position where the sediment-laden flow changes from the surface flow state to the underwater submerged state. Combined with the floating trajectory, the degree of motion separation between the turbid water boundary and the temperature difference boundary, extract the water environment disturbance characteristics of the sediment-laden flow to different dam areas. The water environment disturbance characteristics include the water environment disturbance intensity and the water environment disturbance direction. S4. Based on the structural characteristic data of the dam body, analyze the structurally sensitive direction corresponding to each dam body area, and determine the degree of risk amplification of water environment disturbance by the structural sensitivity of the dam body based on the matching relationship between the structurally sensitive direction and the direction of water environment disturbance. S5. Based on the intensity of water environment disturbance in each dam area and the degree of risk amplification of water environment disturbance by the dam structure sensitivity, analyze the dam structure risk after the water environment disturbance is amplified by the dam structure sensitivity, and generate the water resource environment assessment results of the target reservoir. Step S3 extracts the water environment disturbance characteristics of sediment-laden water flow to different dam areas. Specific steps include: S35. By performing time synchronization and spatial registration on visible light image data, hyperspectral image data and thermal infrared image data at consecutive times, the floating trajectory, sediment-laden water flow boundary and temperature difference boundary are mapped to the same spatial coordinate system of the target reservoir. S36. Construct time series analysis windows centered on each intrusion location, extract the spatial positions of the floating trajectory, sediment-laden water flow boundary, and temperature difference boundary at continuous time intervals within each time series analysis window, and use spline curve fitting to obtain the floating trajectory, sediment-laden water flow boundary migration trajectory, and temperature difference boundary migration trajectory, and determine the first motion separation degree between the floating trajectory and the sediment-laden water flow boundary, the second motion separation degree between the floating trajectory and the temperature difference boundary, and the cooperative migration degree between the sediment-laden water flow boundary and the temperature difference boundary, respectively. Take the direction of the sediment-laden water flow boundary migration trajectory as the direction of water environment disturbance. S37. By weighted summing of the first motion separation degree, the second motion separation degree, and the cooperative migration degree corresponding to each intrusion location, the water environment disturbance characteristic value corresponding to each time moment is obtained, and the average value of the water environment disturbance characteristic value within the corresponding time series analysis window is used as the water environment disturbance intensity of the sediment-laden water flow moving from the surface to the underwater depth at each intrusion location. Step S4, which determines the degree to which the dam structure's sensitivity amplifies the risk of water environment disturbance, includes the following specific steps: S41. Extract the crack extension direction and structural joint extension direction in each dam area based on the dam structure feature data, and use the normal direction of the crack extension direction and structural joint extension direction as the local structural sensitive direction of the corresponding dam area. S42. The absolute value of the cosine of the angle between the sensitive direction of the local structure and the direction of water environment disturbance is taken as the direction matching degree of each dam area. S43. Determine the structural sensitivity of each dam area based on the distribution range of cracks and structural joints within the corresponding dam area. S44. By weighting and summing the structural sensitivity through directional matching degree, we can obtain the degree of risk amplification of the dam structure sensitivity to the corresponding water environment disturbance.

2. The water resource environment assessment method based on machine vision according to claim 1, characterized in that, In step S3, the intrusion location where the sediment-laden water flow transitions from a surface-following state to an underwater intrusion state is determined based on the migration trend of the sediment-laden water flow boundary. Specific steps include: S31. Divide the boundary of the sediment-laden flow into boundary units, and determine the migration velocity, migration direction and boundary curvature change rate of each boundary unit based on the positional changes of each boundary unit at continuous time intervals. S32. Extract the surface water reference velocity and the main direction of the water surface corresponding to the location of each boundary unit based on the floating trajectory. Combine the migration velocity and migration direction of each boundary unit to determine the degree of deviation of the migration velocity and the degree of deviation of the migration direction of each boundary unit. S33. By weighted summing of the migration velocity deviation, migration direction deviation and boundary curvature change rate of each boundary element, the degree of boundary migration anomaly of each boundary element is obtained. S34. Boundary units with a boundary migration anomaly degree greater than the preset boundary migration anomaly threshold are taken as target boundary units. The spatial locations corresponding to the target boundary units are clustered, and the center locations of each cluster are taken as the intrusion locations where the sediment-laden water flow changes from the surface flow state to the underwater intrusion state.

3. The water resource environment assessment method based on machine vision according to claim 2, characterized in that, The specific steps of step S32 include: projecting the migration velocity of each boundary unit onto the corresponding mainstream water direction to obtain the forward migration velocity of each boundary unit, and taking the ratio of the absolute value of the difference between the forward migration velocity and the corresponding surface water reference velocity to the surface water reference velocity as the degree of deviation of the migration velocity of each boundary unit; calculating the deviation angle between the migration direction of each boundary unit and the corresponding mainstream water direction, and taking the ratio of the deviation angle to the average value of all deviation angles as the degree of deviation of the migration direction of each boundary unit.

4. The water resource environment assessment method based on machine vision according to claim 1, characterized in that, The specific steps of step S36 include: S361. Based on the spatial coordinates of the floating trajectory, the sediment-laden flow boundary migration trajectory, and the temperature difference boundary migration trajectory at the same time, calculate the first relative position vector of the sediment-laden flow boundary migration trajectory relative to the floating trajectory, the second relative position vector of the temperature difference boundary migration trajectory relative to the floating trajectory, and the third relative position vector of the temperature difference boundary migration trajectory relative to the sediment-laden flow boundary migration trajectory. S362. Subtract the first relative position vector and the second relative position vector at each time point from the initial relative position vector corresponding to the start time of the time series analysis window to obtain the first relative separation displacement of the sediment-laden water flow boundary relative to the floating trajectory and the second relative separation displacement of the temperature difference boundary relative to the floating trajectory. S363. Calculate the root mean square values ​​of the first relative separation displacement modulus and the second relative separation displacement modulus within the time series analysis window, and calculate the ratio with the cumulative migration length of the floating trajectory within the time series analysis window to obtain the first degree of motion separation between the floating trajectory and the boundary of the sediment-laden water flow and the second degree of motion separation between the floating trajectory and the temperature difference boundary. S364. Subtract the third relative position vector at each time point from the initial third relative position vector corresponding to the start time of the time series analysis window to obtain the relative offset displacement of the temperature difference boundary relative to the sediment-laden water flow boundary, and calculate the root mean square value of the relative offset displacement magnitude within the time series analysis window. S365. Calculate the ratio of the root mean square value to the average cumulative migration length of the sediment-laden flow boundary migration trajectory and the temperature difference boundary migration trajectory to obtain the relative offset between the sediment-laden flow boundary and the temperature difference boundary. Take the reciprocal of the sum of the value 1 and the relative offset as the degree of cooperative migration between the sediment-laden flow boundary and the temperature difference boundary.

5. The water resource environment assessment method based on machine vision according to claim 1, characterized in that, The steps in step S5 to generate the water resources and environmental assessment results for the target reservoir include: S51. The sum of the value 1 and the risk amplification degree of the dam structure sensitivity to water environment disturbance is taken as the risk amplification factor, and the product of the water environment disturbance intensity and the risk amplification factor is taken as the dam structure risk corresponding to each dam area. S52. Output the water environment disturbance characteristics of sediment-laden water flow to different dam areas as the water resource environment assessment results of the target reservoir. When the dam structure risk is greater than the preset dam structure risk threshold, a dam structure risk warning is issued. When the dam structure risk is less than or equal to the preset dam structure risk threshold, no dam structure risk warning is issued.

6. A machine vision-based water resources and environment assessment system, used to implement the machine vision-based water resources and environment assessment method according to any one of claims 1-5, characterized in that, The system includes: The data acquisition module is used to acquire multimodal visual data of the target reservoir and structural feature data of the dam. The multimodal visual data of the water body includes visible light image data, hyperspectral image data and thermal infrared image data. The boundary extraction module is used to extract the floating trajectory of floating objects on the water surface based on visible light image data, analyze the turbidity distribution of water body based on hyperspectral image data and extract the boundary of sediment-laden water flow, and analyze the temperature difference distribution of water body based on thermal infrared image data and extract the temperature difference boundary. The water environment disturbance extraction module is used to determine the intrusion location of the sediment-laden water flow from the surface drift state to the underwater intrusion state based on the migration trend of the sediment-laden water flow boundary. It also extracts the water environment disturbance characteristics of the sediment-laden water flow on different dam areas by combining the floating trajectory, the degree of motion separation between the turbid water boundary and the temperature difference boundary. The water environment disturbance characteristics include the water environment disturbance intensity and the water environment disturbance direction. The dam structure sensitivity analysis module is used to analyze the structural sensitivity direction corresponding to each dam area based on the dam structure characteristic data, and to determine the degree of risk amplification of water environment disturbance by the dam structure sensitivity based on the matching relationship between the structural sensitivity direction and the water environment disturbance direction. The assessment results output module is used to analyze the risk of water environment disturbance amplified by the dam structure sensitivity after the water environment disturbance is amplified, based on the intensity of water environment disturbance in each dam area and the degree of risk amplification of water environment disturbance by the dam structure sensitivity, and to generate the water resource environment assessment results of the target reservoir.

7. An electronic device, comprising: A processor and a memory, wherein the memory stores a computer program that can be called by the processor; characterized in that the processor executes the machine vision-based water resource environment assessment method as described in any one of claims 1-5 by calling the computer program stored in the memory.

8. A computer-readable storage medium, characterized in that, The system stores instructions that, when executed on a computer, cause the computer to perform the machine vision-based water resources and environment assessment method as described in any one of claims 1-5.

Images