Flood situation detection method and device based on big data hyper-fusion and medium

Abstract

The invention discloses a flood situation detection method and device based on big data hyper-fusion and a medium. The method comprises the steps of generating a corresponding ponding area list and a monitoring area list; performing spatial association on the ponding area list, the camera list and the water level gauge sensor list so as to determine the camera and the water level gauge sensor which have a spatial association relationship with the ponding area; for each ponding area, calculating the water depth of the ponding area corresponding to the ponding area according to the water depth data collected by the camera and the water level gauge sensor which have the spatial association relationship with the ponding area, and predicting the predicted ponding depth of the ponding area in the future according to the water depth of the ponding area; dividing camera monitoring grading lists of different grades according to the first camera list, the second camera list and the fourth camera list; and feeding back the gathered flood situation information and the map capturing frequency to a control center, so that the control center carries out rescue scheduling according to the flood situation information, and carries out dynamic scheduling on cameras in the ponding area and the monitoring area.

Description

Technical Field

[0001] The present application relates to the field of big data technology, and specifically to a flood detection method, device and medium based on big data hyper-fusion. Background Art

[0002] As global climate change intensifies, extreme weather events occur frequently, especially natural disasters such as rainstorms and floods, which pose a huge threat to human society. Preventing waterlogging hazards during the flood season and promptly reminding people to evacuate are of great significance to protecting the safety of people and property.

[0003] Traditional flood prevention and rescue systems mostly rely on limited meteorological stations and manual inspections to detect flood conditions. This method has limited data coverage, making it difficult to detect flood conditions in the entire area and issue safety warnings, which brings certain difficulties to flood prevention and rescue. In addition, city-level prevention and control requires a large number of manpower to monitor camera images, which is very costly and will result in problems such as untimely monitoring and poor accuracy. Summary of the invention

[0004] In order to solve the above problems, this application proposes a flood detection method based on big data hyper-fusion, including:

[0005] Acquire waterlogging information of the city to be detected from multiple data sources, generate latitude and longitude dot matrixes corresponding to historical waterlogging areas and predicted waterlogging areas respectively according to the waterlogging information, and merge the latitude and longitude dot matrixes with the waterlogging points collected by the collection equipment to generate corresponding waterlogging area lists and monitoring area lists;

[0006] Obtaining a preset camera list and a water level gauge sensor list, and spatially associating the water accumulation area list with the camera list and the water level gauge sensor list, respectively, to determine the cameras and water level gauge sensors that have a spatial association relationship with the water accumulation area;

[0007] For each waterlogging area, the water depth of the waterlogging area corresponding to the waterlogging area is calculated based on the water depth data collected by the camera and the water level meter sensor that are spatially associated with the waterlogging area, and the predicted water depth of the waterlogging area in the future is predicted based on the water depth of the waterlogging area;

[0008] Determine the first camera list and the second camera list within the coverage of the waterlogging area list and the monitoring area list, and the third camera list that completes the reporting of waterlogging information within a preset time, determine the geographic space to be explored corresponding to the city to be detected and the fourth camera list within the space to be explored according to the buffer zone formed by the third camera list, and divide the camera monitoring hierarchical lists of different levels according to the first camera list, the second camera list and the fourth camera list; wherein cameras at different levels have different capture frequencies;

[0009] The flood information of the relevant waterlogging area is summarized according to the predicted waterlogging depth, and the summarized flood information and the capture frequency are fed back to the control center, so that the control center can carry out rescue dispatch according to the flood information, and dynamically dispatch the cameras inside the waterlogging area and the monitoring area.

[0010] In one implementation of the present application, before calculating the water depth of the water accumulation area corresponding to the water accumulation area based on the water depth data collected by the camera and the water level meter sensor that are spatially associated with the water accumulation area, the method further includes:

[0011] The water depth data collected by the camera at different time points is corrected by the following formula to obtain the corrected water depth data:

[0012]

[0013] Among them, S T represents the water depth data at time point T, λ represents the attenuation rate parameter, T represents the latest time point, t>T-10 Indicates that within ten minutes from the nearest time point, x t Represents the water depth data at time point t.

[0014] In one implementation of the present application, the water depth of the water accumulation area corresponding to the water accumulation area is calculated based on the water depth data collected by the camera and the water level meter sensor that are spatially associated with the water accumulation area, specifically including:

[0015] Determine the observation weight values ​​corresponding to the camera and the water level gauge sensor respectively; wherein the sum of the observation weight value corresponding to the camera and the observation weight value corresponding to the water level gauge sensor is 1;

[0016] Selecting a target camera and a target water level gauge sensor with the largest corresponding water depth data from the cameras and water level gauge sensors that are spatially associated with the waterlogged area;

[0017] The first target water depth data and the second target water depth data corresponding to the target camera and the target water level gauge sensor are obtained respectively, and the first target water depth data and the second target water depth data are weightedly summed according to the observation weight value to obtain the water depth of the water accumulation area corresponding to the water accumulation area.

[0018] In one implementation of the present application, according to the buffer zone formed by the third camera list, the geographic space to be explored corresponding to the city to be detected and the fourth camera list in the space to be explored are determined, and according to the first camera list, the second camera list and the fourth camera list, different levels of camera monitoring hierarchical lists are divided, specifically including:

[0019] From the polygonal fence corresponding to the city to be detected, the areas where the waterlogged area list and the monitored area list are located and the buffer zone formed by the third camera list are screened out to obtain the geographic space to be explored;

[0020] Selecting any one camera from each grid unit of the geographic space to be explored to generate a corresponding fourth camera list according to the camera;

[0021] splicing the first camera list, the second camera list and the fourth camera list to obtain a spliced ​​target camera list;

[0022] A truncation ratio for grading the target camera list is determined, and the target camera list is truncated into a plurality of camera monitoring grading lists of different levels according to the truncation ratio.

[0023] In one implementation of the present application, determining a cutoff ratio for grading the target camera list specifically includes:

[0024] An objective function and constraints for optimizing the truncation ratio and the screenshot ratio are constructed; the objective function is expressed as:

[0025] maxω1*p1*f1+ω2*p2*f2+ω3*p3*f3

[0026] Among them, ω1, ω2, and ω3 represent the relative importance weights of different levels, p1, p2, and p3 represent the cutoff ratios corresponding to different levels, and f1, f2, and f3 represent the capture frequencies of different levels;

[0027] According to the constraint conditions and the objective function, the cutoff ratio for grading the target camera list and the capture frequencies corresponding to the graded camera monitoring lists of each grade are determined.

[0028] In one implementation of the present application, the latitude and longitude dot matrix corresponding to the historical waterlogging area and the predicted waterlogging area is generated according to the waterlogging information, specifically including:

[0029] Based on the S2 geocoding grid coverage system, determine the boundaries of the historical waterlogging area and the predicted waterlogging area to be covered, and determine the initial grid unit level;

[0030] Starting from the initial grid unit level, the grid unit level is successively reduced until the grid unit can cover the historical waterlogging area and the predicted waterlogging area, thereby obtaining the corresponding target grid unit level;

[0031] According to the target grid unit level, the grid units covered by the historical waterlogging area and the predicted waterlogging area are converted into corresponding longitude and latitude coordinates to obtain a longitude and latitude dot matrix composed of the longitude and latitude coordinates.

[0032] In one implementation of the present application, the latitude and longitude dot matrix and the waterlogging points collected by the collection device are merged to generate a corresponding waterlogging area list and a monitoring area list, specifically including:

[0033] The latitude and longitude dot matrix and the water accumulation points collected by the collection device are combined to generate a coordinate array consisting of a plurality of coordinate points;

[0034] Clustering the coordinate array by a preset clustering algorithm to obtain a plurality of cluster groups, and filtering out the specified cluster groups whose sample numbers in the cluster groups do not meet the minimum sample number;

[0035] For each cluster group after filtering, determining whether the number of samples in the cluster group is not less than a preset number of small samples;

[0036] If yes, generate the corresponding region boundary for each coordinate point included in the cluster group according to the convex hull algorithm; if no, generate the region boundary corresponding to the cluster group according to the buffer method;

[0037] The area where the regional boundary is located is taken as a waterlogging area or a monitoring area to obtain a corresponding waterlogging area list and a monitoring area list.

[0038] In one implementation of the present application, predicting the predicted water depth of the waterlogged area in the future according to the water depth of the waterlogged area specifically includes:

[0039] Based on the B function, a depth prediction model is constructed to predict the future water depth;

[0040] The depth prediction model is expressed as:

[0041]

[0042] in, is the predicted water depth after k minutes in the future, A is the amplitude parameter, b is the baseline level, Z is the future Z minutes, L is the historical L minutes, α and β are both shape parameters;

[0043] The depth prediction model is used to predict the future water depth of the waterlogged area.

[0044] The embodiment of the present application provides a flood detection device based on big data hyper-fusion, the device comprising:

[0045] at least one processor;

[0046] and, a memory communicatively coupled to the at least one processor;

[0047] Among them, the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a flood detection method based on big data hyper-fusion as described in any of the above items.

[0048] The embodiment of the present application provides a non-volatile computer storage medium storing computer executable instructions, wherein the computer executable instructions are configured as follows:

[0049] A flood detection method based on big data hyper-fusion as described in any of the above items.

[0050] The flood detection method based on big data hyper-fusion proposed in this application can bring the following

[0051] Beneficial effects:

[0052] The waterlogging information of the city to be tested is obtained from multiple data sources, and then the corresponding waterlogging area list and monitoring area list are generated. This breaks the limitations of the traditional flood prevention and rescue system that relies on limited meteorological stations and manual inspections, significantly improves the data coverage and accuracy of flood detection, and enables the flood situation in the entire region to be fully and real-time grasped, providing strong support for safety warnings.

[0053] By spatially correlating the camera list, water level gauge sensor list and water accumulation area list, the cameras and water level gauge sensors that are spatially correlated with the water accumulation area can be accurately determined, which improves the pertinence and efficiency of data collection and makes the calculation of the water depth of the water accumulation area and the predicted water accumulation depth more accurate and reliable.

[0054] By dividing the camera monitoring list into different levels and setting different capture frequencies according to the different levels of cameras, effective management and dynamic scheduling of camera resources are achieved, and city-wide flood detection is achieved with minimal cost. At the same time, it can also effectively save the city’s camera bandwidth and computing power. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0056] Figure 1 A schematic diagram of a flood detection system architecture based on big data fusion provided in an embodiment of the present application;

[0057] Figure 2 A schematic diagram of a flood detection method based on big data hyper-fusion provided in an embodiment of the present application;

[0058] Figure 3 A schematic diagram of the structure of a flood detection device based on big data hyper-fusion provided in an embodiment of the present application. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0060] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0061] like Figure 1The schematic diagram of the flood detection system architecture based on big data fusion is shown in the figure. The system includes peripheral systems, access layers, hyper-fusion layers and display layers. Among them, the peripheral system includes data sources such as cameras, sensors and predicted rainfall. Through the above data sources, the city's water accumulation information can be collected, including camera water depth, number of people and vehicles data, water level gauge water depth data, and grid predicted water depth data predicted by the waterlogging simulation model. After the water accumulation information enters the system through the access layer, it is processed by the hyper-fusion layer. The processed data includes two types. One is the display type, such as the water accumulation area table, the water accumulation area-camera relationship table, the water accumulation area-water depth table, etc. These output results are finally presented in the display part as a list of water accumulation areas (water accumulation point name, current water depth, predicted increment, number of people and vehicles). Through the water accumulation area table, it is possible to count the number of dangerous situations in the city in real time, sort the emergency situations according to the current water depth and number of people and vehicles, and clarify the rescue priority of each water accumulation area. The second is the scheduling type, that is, the camera monitoring hierarchical list. The camera monitoring hierarchical list is a configuration table for the camera recognition service in the peripheral system, indicating which cameras need to be captured and visually recognized at present, as well as the corresponding capture frequency of the cameras. Through the camera monitoring extension list, necessary areas can be monitored to support emergency dispatch, while reducing the monitoring of unnecessary areas, saving the city camera network bandwidth and computing power.

[0062] The data processing process of the hyper-convergence layer is the core content for realizing flood detection. The embodiment of the present application is based on the hyper-convergence layer and provides a flood detection method based on big data hyper-convergence. It evaluates and predicts urban waterlogging conditions by processing multi-source waterlogging information in real time, and ultimately provides users with comprehensive waterlogging information, providing data support for urban flood control and traffic management.

[0063] like Figure 2 As shown, a flood situation detection method based on big data hyper-fusion provided in an embodiment of the present application includes:

[0064] S201: Obtain waterlogging information of the city to be inspected from multiple data sources, generate latitude and longitude dot matrices corresponding to historical waterlogging areas and predicted waterlogging areas according to the waterlogging information, and merge the latitude and longitude dot matrices with the waterlogging points collected by the collection equipment to generate corresponding waterlogging area lists and monitoring area lists.

[0065] Data sources include cameras, water level gauge sensors, manually configured historical waterlogging area lists, and predicted waterlogging grid lists predicted by waterlogging simulation models. The static waterlogging area list is a predefined historical waterlogging area used to assist in identifying new waterlogging areas or correcting existing waterlogging areas. The predicted waterlogging grid list refers to the waterlogging prediction model that regularly simulates waterlogging for the entire city in the next 24 hours based on weather data. Waterlogging information of the city to be tested is obtained from the above multiple data sources. The waterlogging information includes the latitude and longitude coordinates and water depth data of the waterlogging point. After converting it to a unified format, the latitude and longitude dot matrix corresponding to the historical waterlogging area and the predicted waterlogging area can be generated accordingly.

[0066] Among them, the latitude and longitude dot matrix is ​​generated by setting the appropriate S2 Cell level through the S2 geocoding grid coverage system. The S2 Cell level, that is, the grid cell level, determines the size and density of the grid. When determining the appropriate grid cell level, it is first necessary to determine the boundaries of the historical waterlogging area and the predicted waterlogging area to be covered based on the S2 geocoding grid coverage system (usually polygonal or circular), and determine the initial grid cell level. The initial grid cell level is a higher-level grid single level obtained by recursive search. Starting from the initial grid cell level, the grid cell level is lowered in sequence. Each time the grid cell level is refined, smaller grid cells are gradually determined until the grid cell can cover the historical waterlogging area and the predicted waterlogging area, and the corresponding target grid cell level is obtained. S2 coding also has the characteristics of efficient longitude and latitude calculation. Each grid cell has a unique integer ID, which represents its position and level in the spherical subdivision structure. Using these IDs, the longitude and latitude coordinates of the center point and vertex of each grid cell can be efficiently calculated. Through the above characteristics, when generating the coverage grid according to the target grid unit level, the coverage grid cells of the historical waterlogging area and the predicted waterlogging area can be quickly converted into specific longitude and latitude coordinates in the geographic coordinate system, thereby forming a longitude and latitude dot matrix.

[0067] After obtaining the latitude and longitude dot matrix, the latitude and longitude dot matrix and the waterlogging points collected by the acquisition device are merged to generate the corresponding waterlogging area list and monitoring area list. Specifically, the latitude and longitude dot matrix and the waterlogging points collected by the acquisition device are merged to generate a coordinate array consisting of several coordinate points. The coordinate array contains all possible waterlogging points and predicted waterlogging points, and each corresponding area boundary can be generated through the coordinate array.

[0068] First, the coordinate array is clustered by a preset clustering algorithm to obtain multiple cluster groups. The clustering algorithm can select the DBSCAN algorithm, and each cluster group obtained by clustering has a corresponding cluster label. For the coordinate point with a label of -1, it will be identified as noise and filtered out. For the cluster group after clustering, filter out the specified cluster group whose sample number in the cluster group does not meet the minimum sample number, and avoid identifying noise data or too small areas as effective waterlogging areas. For each cluster group after filtering, it is necessary to calculate its boundary. The embodiment of the present application provides two boundary generation methods. The specific method needs to be selected by whether the sample number in the cluster group is not less than the preset small sample number. If it is not less than the small sample number, it is necessary to generate the minimum outer polygon for each coordinate point contained in the cluster group according to the convex hull algorithm, and the minimum outer polygon is the regional boundary corresponding to the cluster group. If it is less than the small sample number, it is necessary to generate the minimum enclosing circular area corresponding to the coordinate point contained in the cluster group by the buffer method, and the minimum enclosing circular area is the final regional boundary. Through the above method, corresponding regional boundaries can be generated for waterlogging information from different data sources. The area enclosed by the regional boundaries is the waterlogging area or the monitoring area. By integrating other information of the waterlogging area or the monitoring area, the corresponding waterlogging area list and monitoring area list can be obtained.

[0069] In the above process, the waterlogging area is obtained by integrating the data that confirms that there is waterlogging. When generating the waterlogging area list, the data source that needs to be selected is the data source that collects historical waterlogging data, including cameras, water level gauge sensors, and manually configured historical waterlogging area lists; the monitoring area is based on the waterlogging area, and then integrates the predicted waterlogging grid list predicted by the waterlogging simulation model. In other words, the waterlogging area list covers all the waterlogging points that have already experienced waterlogging. In addition to the waterlogging points that have been detected, the monitoring area list also includes some waterlogging points that are predicted to have waterlogging but have not yet experienced waterlogging.

[0070] It should be noted that the purpose of distinguishing between waterlogging areas and monitoring areas is that the waterlogging areas allow downstream users to aggregate a large amount of scattered camera and water level meter information data. The monitoring area is to select cameras that need to enable visual recognition models and reduce the number of cameras with low probability of waterlogging, thereby saving computing power and network bandwidth resources.

[0071] S202: Obtain a preset camera list and water level gauge sensor list, and spatially associate the water accumulation area list with the camera list and the water level gauge sensor list to determine the cameras and water level gauge sensors that are spatially associated with the water accumulation area.

[0072] Get the preset camera list and water level sensor list, which contain the location information of the cameras and water level sensors. Through the method of spatial association analysis, spatial association is performed on the list of waterlogged areas, the list of cameras, and the list of water level sensors, so as to determine which cameras and water level sensors are close to or overlap with the locations of waterlogged areas in space, thereby determining the spatial association relationship between them. By identifying water level sensors and cameras that are spatially associated with waterlogged areas, the waterlogging situation can be assessed more accurately, potential flood risks can be predicted, and a scientific basis can be provided for flood prevention and disaster reduction.

[0073] When performing spatial association analysis, the S2 code corresponding to each coordinate point in the above list is generated by S2 coding. By associating the list of waterlogged areas with the list of cameras and the list of water level gauge sensors through S2 coding, the cameras and water level gauge sensors that are close to or around each waterlogged area can be obtained. The periphery here refers to the coverage of the grid unit where the waterlogged area is located. As long as the cameras and water level gauge sensors are located in this coverage, they can be regarded as having a spatial association relationship with the waterlogged area.

[0074] S203: For each waterlogging area, the water depth of the waterlogging area corresponding to the waterlogging area is calculated based on the water depth data collected by the camera and water level meter sensor that are spatially associated with the waterlogging area, and the predicted water depth of the waterlogging area in the future is predicted based on the water depth of the waterlogging area.

[0075] For each flooded area, the on-site visual information provided by the cameras that have a spatial correlation with it and the real-time water depth data collected by the water level sensor are integrated, and the actual water depth of the current flooded area is accurately calculated through comprehensive analysis of these data. On this basis, the future flooded depth and the trend of flooded depth changes in the flooded area are further predicted, so as to provide timely and accurate flooding assessment and early warning information to relevant departments so that they can take effective disaster prevention and mitigation measures.

[0076] In one embodiment, since the accuracy of water depth recognition by the camera is not very stable and will fluctuate with the disturbance of the picture, the water depth data collected by the camera needs to be corrected. Specifically, it can be corrected by the following formula:

[0077]

[0078] Among them, S T represents the water depth data at time point T, λ represents the attenuation rate parameter, T represents the latest time point, t>T-10 Indicates that within ten minutes from the nearest time point, x t Represents the water depth data at time point t.

[0079] According to the corrected camera water depth data, combined with the water depth data collected by the water level gauge sensor, the water depth of the waterlogged area can be calculated. In this process, the observation weight values ​​corresponding to the camera and the water level gauge sensor must be determined first. The observation weight value can be obtained based on the analysis of historical waterlogging data, and the sum of the observation weight value corresponding to the camera and the observation weight value corresponding to the water level gauge sensor is 1. Since there may be multiple spatially associated cameras and water level gauge sensors in each waterlogged area, the actual water depth of the waterlogged area needs to be determined by the maximum water depth collected. Therefore, it is necessary to select the target camera and target water level gauge sensor with the largest corresponding water depth data from the cameras and water level gauge sensors that have a spatial association relationship with the waterlogged area, and obtain the first target water depth data and the second target water depth data corresponding to the target camera and the target water level gauge sensor, respectively. In this way, the first target water depth data and the second target water depth data are weighted and summed according to the observation weight value, and the water depth of the waterlogged area corresponding to the waterlogged area can be obtained.

[0080] The water depth of the waterlogged area can be calculated using the following formula:

[0081]

[0082] in, is the depth of the waterlogged area at time point T, W T is the water depth data collected by the water level sensor at time point T, a represents the observation weight corresponding to the water level sensor, max(S T ) is the first target water depth data, max(W T ) represents the second target water depth data.

[0083] Each time the current water depth of the waterlogged area is calculated, it is saved in a persistent memory. The water depth {dt} of the waterlogged area recorded in the history of L minutes is retrieved, where t = 1, 2, ..., L. The predicted water depth of the waterlogged area in the future Z minutes is calculated. This process can be achieved through the preset B function. Through the preset B function (Beta function), a depth prediction model for predicting the future water depth is constructed. The depth prediction model can be expressed as:

[0084]

[0085] in, is the predicted water depth after k minutes in the future, A is the amplitude parameter, b is the baseline level, Z is the future Z minutes, L is the historical L minutes, and α and β are shape parameters.

[0086] For the above depth prediction model, A, b, α and β can be estimated by fitting historical data with the least squares method. After clarifying the above parameters, the predicted water depth of the corresponding waterlogged area in the future can be obtained based on the water depth of the waterlogged area within the historical L minutes. It should be noted that if α<β, the curve is biased to the left, indicating that the waterlogging is in the rising stage. If α>β, the curve is biased to the right, indicating that the waterlogging is in the falling stage. If α≈β, the curve is symmetrical and may be close to the peak.

[0087] S204: Determine the first camera list and the second camera list within the coverage of the waterlogging area list and the monitoring area list, and the third camera list that completes reporting of waterlogging information within a preset time, and determine the fourth camera list within the geographic space to be explored corresponding to the city to be detected and the space to be explored according to the buffer zone formed by the third camera list, and divide the camera monitoring graded lists of different levels according to the first camera list, the second camera list and the fourth camera list; wherein cameras at different levels have different capture frequencies.

[0088] The camera recognition service is a large-scale system that can capture and visually recognize hundreds of thousands of cameras in a city. However, due to the limited bandwidth of the city camera network, the visual recognition algorithm is a computing-intensive task, and the number and frequency of cameras for capturing cannot be set indefinitely. Therefore, it is also necessary to dynamically schedule and plan the city cameras based on the first camera list and the second camera list within the coverage of the flooded area list and the monitoring area list, as well as the third camera list that completes the reporting of flooded information within the preset time. By classifying and grading the cameras, different capture frequencies are assigned to cameras of different levels, so as to reduce the visual recognition computing power as much as possible while effectively ensuring the city's ability to perceive flood risks.

[0089] Specifically, from the polygonal fence corresponding to the city to be detected, the areas where the waterlogged area list and the monitoring area list are located and the buffer zone formed by the third camera list are screened out to obtain the geographic space to be explored. The buffer zone here refers to the polygonal area formed by 100 meters around the coordinates of the third camera. Any camera is selected from each grid unit of the geographic space to be explored, and the fourth camera list can be generated accordingly based on the selected camera.

[0090] Furthermore, the cameras in the first camera list and the second camera list are arranged in reverse order according to the most recently reported water depth, and the first camera list, the second camera list and the fourth camera list are spliced ​​in sequence to obtain a spliced ​​target camera list.

[0091] Further, determine the truncation ratio for grading the target camera list, and truncate the target camera list into camera monitoring grading lists of multiple different levels according to the truncation ratio. It should be noted that when truncating the list, it needs to be truncated according to lengths such as MAX_LENGTH*p1, MAX_LENGTH*p2, MAX_LENGTH*p3, where p1, p2, and p3 are the truncation ratios corresponding to the three levels respectively (0 ≤ p i ≤ 1, and p1 + p2 + p3 = 1), and MAX_LENGTH represents the maximum total length of the camera list of the target camera list. For each level's camera monitoring grading list, the capture frequency of the cameras is different, which can be expressed as f1 images per second, f2 images per second, and f3 images per second respectively.

[0092] In the embodiments of the present application, the truncation ratio and the capture frequency need to find corresponding optimal solutions through an optimization algorithm. First, construct an objective function and constraint conditions for optimizing the truncation ratio and the capture ratio.

[0093] The objective function is expressed as:

[0094] max ω1*p1*f1 + ω2*p2*f2 + ω3*p3*f3

[0095] Where ω1, ω2, and ω3 represent the relative importance weights of different levels, p1, p2, and p3 represent the truncation ratios corresponding to different levels respectively, and f1, f2, and f3 represent the capture frequencies of different levels.

[0096] The constraint conditions include:

[0097] The relative importance weight of each level: ω1 + ω2 + ω3 = 1; computing power constraint: (p1*f1 + p2*f2 + p3*f3)*MAX_LENGTH*VC ≤ CP; sum of ratios is 1: p1 + p2 + p3 = 1; specific ratio requirement: p1 > number of necessary for level 1 / MAX_LENGTH; specific ratio requirement: p2 > number of necessary for level 2 / MAX_LENGTH; specific ratio requirement: p3 > number of necessary for level 3 / MAX_LENGTH; specific frequency requirement: f1 < 1 / 5; specific frequency requirement: f2 < 1 / 60; specific frequency requirement: f2 < 1 / 1200; all variables are non - negative: p i ≥ 0.001, f i > 0.001 (i = 1, 2, 3); bandwidth constraint: MAX_LENGTH < BD / bd. Where CP is the total available computing power, VC is the computing power consumption of single - image visual recognition, BD is the available bandwidth, and bd is the bandwidth occupied by a single camera.

[0098] According to the above constraints and objective function, the optimal cut-off ratio for grading the target camera list and the corresponding capture frequency of each level of camera monitoring grading list can be found.

[0099] S205: Summarize the flood information of the relevant flooded areas according to the predicted water depth, and feed back the summarized flood information and the capture frequency to the control center, so that the control center can dispatch rescue operations according to the flood information, and dynamically dispatch the cameras in the flooded areas and the monitoring areas.

[0100] The water depth data corresponding to each waterlogged area is summarized to obtain the flood information of the relevant waterlogged areas, which records the actual conditions of each waterlogged area in detail. Subsequently, the summarized flood information and the capture frequency of the corresponding area are fed back to the control center. Based on the flood information received, the control center counts the flood dangers in the city, identifies the waterlogged areas that need rescue the most, and quickly and accurately makes rescue dispatch decisions, including dispatching rescue teams and deploying rescue materials to deal with possible flood disasters. At the same time, the control center can also dynamically dispatch cameras in the waterlogged area and the wider monitoring area based on this information, and reduce monitoring of unnecessary areas by adjusting the camera's capture frequency, saving network bandwidth and computing power.

[0101] The above are embodiments of the method proposed in this application. Based on the same idea, some embodiments of this application also provide devices and non-volatile computer storage media corresponding to the above methods.

[0102] Figure 3 A structural diagram of a flood detection device based on big data hyper-fusion provided in an embodiment of the present application. Figure 3 As shown, including:

[0103] at least one processor; and,

[0104] at least one processor is communicatively connected to a memory; wherein,

[0105] The memory stores instructions that can be executed by at least one processor, and the instructions are executed by at least one processor so that the at least one processor executes a flood detection method based on big data hyper-fusion as described in any of the above items.

[0106] The embodiment of the present application provides a non-volatile computer storage medium storing computer executable instructions, wherein the computer executable instructions are configured as follows:

[0107] A flood detection method based on big data hyper-fusion as described in any of the above items.

[0108] Each embodiment in this application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device and medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0109] The devices and media provided in the embodiments of the present application correspond one-to-one to the methods. Therefore, the devices and media also have similar beneficial technical effects as the corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the devices and media will not be repeated here.

[0110] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0111] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0112] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0113] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0114] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0115] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0116] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0117] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0118] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A flood detection method based on big data hyper-fusion, characterized in that: The method comprises: Acquire waterlogging information of the city to be detected from multiple data sources, generate latitude and longitude dot matrixes corresponding to historical waterlogging areas and predicted waterlogging areas respectively according to the waterlogging information, and merge the latitude and longitude dot matrixes with the waterlogging points collected by the collection equipment to generate corresponding waterlogging area lists and monitoring area lists; Obtaining a preset camera list and a water level gauge sensor list, and spatially associating the water accumulation area list with the camera list and the water level gauge sensor list, respectively, to determine the cameras and water level gauge sensors that have a spatial association relationship with the water accumulation area; For each waterlogging area, the water depth of the waterlogging area corresponding to the waterlogging area is calculated based on the water depth data collected by the camera and the water level meter sensor that are spatially associated with the waterlogging area, and the predicted water depth of the waterlogging area in the future is predicted based on the water depth of the waterlogging area; Determine the first camera list and the second camera list within the coverage of the waterlogging area list and the monitoring area list, and the third camera list that completes the reporting of waterlogging information within a preset time, determine the geographic space to be explored corresponding to the city to be detected and the fourth camera list within the space to be explored according to the buffer zone formed by the third camera list, and divide the camera monitoring hierarchical lists of different levels according to the first camera list, the second camera list and the fourth camera list; wherein cameras at different levels have different capture frequencies; The flood information of the relevant waterlogging area is summarized according to the predicted waterlogging depth, and the summarized flood information and the capture frequency are fed back to the control center, so that the control center can carry out rescue dispatch according to the flood information, and dynamically dispatch the cameras inside the waterlogging area and the monitoring area.

2. A flood detection method based on big data hyper-fusion according to claim 1, characterized in that: Before calculating the water depth of the water accumulation area corresponding to the water accumulation area according to the water depth data collected by the camera and the water level sensor that are spatially associated with the water accumulation area, the method further includes: The water depth data collected by the camera at different time points is corrected by the following formula to obtain the corrected water depth data: Among them, S T represents the water depth data at time point T, λ represents the attenuation rate parameter, T represents the latest time point, t>T-10 Indicates that within ten minutes from the nearest time point, x t Represents the water depth data at time point t.

3. A flood detection method based on big data hyper-fusion according to claim 2, characterized in that: Calculating the water depth of the water accumulation area corresponding to the water accumulation area according to the water depth data collected by the camera and the water level meter sensor that are spatially associated with the water accumulation area includes: Determine the observation weight values ​​corresponding to the camera and the water level gauge sensor respectively; wherein the sum of the observation weight value corresponding to the camera and the observation weight value corresponding to the water level gauge sensor is 1; Selecting a target camera and a target water level gauge sensor with the largest corresponding water depth data from the cameras and water level gauge sensors that are spatially associated with the waterlogged area; The first target water depth data and the second target water depth data corresponding to the target camera and the target water level gauge sensor are obtained respectively, and the first target water depth data and the second target water depth data are weightedly summed according to the observation weight value to obtain the water depth of the water accumulation area corresponding to the water accumulation area.

4. A flood situation detection method based on big data hyper-fusion according to claim 1, characterized in that: According to the buffer zone formed by the third camera list, the geographic space to be explored corresponding to the city to be detected and the fourth camera list in the space to be explored are determined, and according to the first camera list, the second camera list and the fourth camera list, different levels of camera monitoring hierarchical lists are divided, specifically including: From the polygonal fence corresponding to the city to be detected, the areas where the waterlogged area list and the monitored area list are located and the buffer zone formed by the third camera list are screened out to obtain the geographic space to be explored; Selecting any one camera from each grid unit of the geographic space to be explored to generate a corresponding fourth camera list according to the camera; splicing the first camera list, the second camera list and the fourth camera list to obtain a spliced ​​target camera list; A truncation ratio for grading the target camera list is determined, and the target camera list is truncated into a plurality of camera monitoring grading lists of different levels according to the truncation ratio.

5. A flood detection method based on big data hyper-fusion according to claim 4, characterized in that: Determining a cutoff ratio for grading the target camera list, specifically comprising: An objective function and constraints for optimizing the truncation ratio and the screenshot ratio are constructed; the objective function is expressed as: maxω1*p1*f1+ω2*p2*f2+ω3*p3*f3 Among them, ω1, ω2, and ω3 represent the relative importance weights of different levels, p1, p2, and p3 represent the cutoff ratios corresponding to different levels, and f1, f2, and f3 represent the capture frequencies of different levels; According to the constraint conditions and the objective function, the cutoff ratio for grading the target camera list and the capture frequencies corresponding to the graded camera monitoring lists of each grade are determined.

6. The flood situation detection method based on big data hyper-fusion according to claim 1 is characterized in that: Generating latitude and longitude dot matrices corresponding to the historical waterlogging area and the predicted waterlogging area respectively according to the waterlogging information, specifically including: Based on the S2 geocoding grid coverage system, determine the boundaries of the historical waterlogging area and the predicted waterlogging area to be covered, and determine the initial grid unit level; Starting from the initial grid unit level, the grid unit level is successively reduced until the grid unit can cover the historical waterlogging area and the predicted waterlogging area, thereby obtaining the corresponding target grid unit level; According to the target grid unit level, the grid units covered by the historical waterlogging area and the predicted waterlogging area are converted into corresponding longitude and latitude coordinates to obtain a longitude and latitude dot matrix composed of the longitude and latitude coordinates.

7. A flood situation detection method based on big data hyper-fusion according to claim 6, characterized in that: The latitude and longitude dot matrix and the waterlogging points collected by the collection device are combined to generate a corresponding waterlogging area list and a monitoring area list, specifically including: The latitude and longitude dot matrix and the water accumulation points collected by the collection device are combined to generate a coordinate array consisting of a plurality of coordinate points; Clustering the coordinate array by a preset clustering algorithm to obtain a plurality of cluster groups, and filtering out the specified cluster groups whose sample numbers in the cluster groups do not meet the minimum sample number; For each cluster group after filtering, determining whether the number of samples in the cluster group is not less than a preset number of small samples; If yes, generate the corresponding region boundary for each coordinate point included in the cluster group according to the convex hull algorithm; if no, generate the region boundary corresponding to the cluster group according to the buffer method; The area where the regional boundary is located is taken as a waterlogging area or a monitoring area to obtain a corresponding waterlogging area list and a monitoring area list.

8. The flood situation detection method based on big data hyper-fusion according to claim 1 is characterized in that: Predicting the predicted water depth of the waterlogged area in the future according to the water depth of the waterlogged area specifically includes: Based on the B function, a depth prediction model is constructed to predict the future water depth; The depth prediction model is expressed as: in, is the predicted water depth after k minutes in the future, A is the amplitude parameter, b is the baseline level, Z is the future Z minutes, L is the historical L minutes, α and β are both shape parameters; The depth prediction model is used to predict the future water depth of the waterlogged area.

9. A flood detection device based on big data hyper-fusion, characterized in that: The device comprises: at least one processor; and, a memory communicatively coupled to the at least one processor; The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a flood detection method based on big data hyper-fusion as described in any one of claims 1-8.

10. A non-volatile computer storage medium storing computer executable instructions, characterized in that: The computer executable instructions are configured to: A flood detection method based on big data hyper-fusion as described in any one of claims 1 to 8.

Images