A method and system for arranging hydrological detection equipment based on the Internet of Things

Abstract

The present invention relates to a method and system for layout of hydrological detection equipment based on the Internet of Things, belonging to the field of hydrological monitoring technology, obtaining multiple frames of continuous image information of a target detection area within a preset time period, and introducing a hidden Markov model and a Baum-Welch algorithm for analysis to generate a time series dynamic model of the target detection area; based on the time series dynamic model, the probability of occurrence of a hydrological natural disaster type is calculated in a Bayesian network algorithm and analyzed and planned to obtain a first installation area; a hydrological detection equipment simulation model is constructed, and the hydrological detection equipment simulation model is imported into the first installation area for simulated measurement and screening of the installation area to obtain a second installation area; a quadratic polynomial regression curve model is constructed, and the final layout point screening of the second installation area is performed based on the model to obtain a final layout plan. The present invention can screen and plan the measurement layout points of the hydrological detection equipment, achieve accurate and efficient measurement effects, and has high reliability.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrological monitoring, and in particular to a hydrological detection equipment layout method and system based on the Internet of Things. Background Art

[0002] Hydrological detection equipment is an instrument or device for monitoring and measuring hydrological parameters, and is widely used in monitoring hydrological variables, water resources management, and environmental protection. The traditional method of measuring hydrological data usually involves installing hydrological detection equipment in the target area to be measured, and transmitting the monitored and measured data to fixed receiving points such as hydrological monitoring stations. In order to ensure the accuracy and reliability of the monitoring data, the selection of the layout points of the hydrological detection equipment is crucial; however, the installation points selected by the current hydrological detection equipment layout method are difficult to accurately select according to the terrain requirements, making it impossible for the hydrological detection equipment to accurately and quickly measure the target hydrological data, and it is easy to have poor signal transmission quality or interruption, resulting in the inability to read and receive hydrological monitoring data. At the same time, it will cause errors in the monitored data, making it difficult to achieve accurate measurement results. Therefore, an accurate and efficient hydrological detection equipment layout method is needed to solve the above problems. Summary of the invention

[0003] The present invention overcomes the deficiencies of the prior art and provides a method and system for arranging hydrological detection equipment based on the Internet of Things.

[0004] To achieve the above object, the technical solution adopted by the present invention is:

[0005] The first aspect of the present invention provides a method for arranging hydrological detection equipment based on the Internet of Things, comprising the following steps:

[0006] Acquire multiple frames of continuous image information of the target detection area within a preset time period, perform feature extraction on the multiple frames of continuous image information to obtain multiple continuous feature vectors, and introduce a hidden Markov model to analyze the multiple continuous feature vectors to generate a time series dynamic model of the target detection area;

[0007] Based on the time series dynamic model, several groups of historical water flow characteristic data are obtained, and the occurrence probability of a hydrological natural disaster type is calculated in a Bayesian network algorithm according to each group of the historical water flow characteristic data, and analysis and planning are performed based on the occurrence probability to obtain a first installation area;

[0008] Constructing a hydrological detection equipment simulation model, importing the hydrological detection equipment simulation model into the first installation area for simulation measurement, obtaining simulation signal transmission frequency band results and signal transmission interruption results, and screening the installation area according to the simulation signal transmission frequency band results and signal transmission interruption results to obtain a second installation area;

[0009] By calculating the second installation area, a point weight descending list is constructed, and a quadratic polynomial regression curve model is established based on the actual water flow characteristic data and the historical water flow characteristic data. The quadratic polynomial regression curve model is analyzed to match the final layout points in the point weight descending list to obtain the final layout plan.

[0010] Furthermore, in a preferred embodiment of the present invention, the method of acquiring multiple frames of continuous image information of a target detection area within a preset time period, performing feature extraction on the multiple frames of the continuous image information to obtain multiple continuous feature vectors, and introducing a hidden Markov model to analyze the multiple continuous feature vectors to generate a time series dynamic model of the target detection area specifically includes the following steps:

[0011] The high-definition camera on the drone is controlled by the Internet of Things technology to continuously capture and shoot the target detection area, and obtain multiple frames of continuous image information of the target detection area within a preset time period;

[0012] Based on the HOG algorithm, feature extraction is performed on the continuous image information of multiple frames, each frame of image information is divided into a number of area blocks, a Sobel filter is added to the HOG algorithm to calculate the gradient size and gradient direction of the pixel points in each area block, a gradient histogram is constructed according to the gradient size and the gradient direction, and the gradient histograms constructed by the multiple frames of image information are connected to obtain multiple continuous feature vectors;

[0013] Introducing a hidden Markov model to analyze the plurality of continuous feature vectors, determining a target hidden state and a target observed variable based on each feature vector, constructing hidden states based on the target hidden state and the target observed variable, respectively, to obtain a plurality of dynamic feature sequences;

[0014] The hidden state sequence of each dynamic feature sequence is analyzed and calculated based on the Baum-Welch algorithm to obtain dynamic model parameters, and a temporal dynamic model of the target detection area is generated based on the dynamic model parameters.

[0015] Furthermore, in a preferred embodiment of the present invention, the method of obtaining several groups of historical water flow characteristic data based on the time series dynamic model, calculating the occurrence probability of a hydrological natural disaster type in a Bayesian network algorithm according to each group of the historical water flow characteristic data, and performing analysis and planning based on the occurrence probability to obtain a first installation area specifically includes the following steps:

[0016] Acquire geographical terrain information of the target detection area, and search for hydrological and natural disasters in a big data network based on the geographical terrain information to obtain a plurality of associated hydrological and natural disaster types;

[0017] Acquire several groups of historical water flow characteristic data based on the time series dynamic model of the target detection area; wherein the water flow characteristics include flow velocity, flow direction, flow rate and increase;

[0018] Constructing variable nodes and variable dependency relationships in a Bayesian network based on each group of the historical water flow characteristics and a plurality of the hydrological natural disaster types, assigning a probability to each variable node according to the variable dependency relationship, and obtaining a probability distribution of each variable node;

[0019] The state transition probability between each set of historical water flow characteristic data and each associated hydrological natural disaster type is determined through the probability distribution of each variable node, and finally the occurrence probability corresponding to each hydrological natural disaster type is determined according to the state transition probability;

[0020] Extract all types of hydrological natural disasters with a probability of occurrence greater than a preset probability of occurrence, generate search tags according to all types of hydrological natural disasters with a probability of occurrence greater than a preset probability of occurrence, and search in a big data network based on the search tags to obtain N hydrological measurement indicators;

[0021] A terrain distribution model is obtained according to a time series dynamic model of the target detection area, and the terrain distribution model is analyzed and planned based on N hydrological measurement indicators to obtain a first installation area.

[0022] Furthermore, in a preferred embodiment of the present invention, the terrain distribution model is obtained according to the time series dynamic model of the target detection area, and the terrain distribution model is analyzed and planned based on N hydrological measurement indicators to obtain the first installation area, which specifically includes the following steps:

[0023] Acquire a terrain type of a target detection area, separate a sub-model from a time series dynamic model of the target detection area to obtain a terrain distribution model, extract all terrain features based on the terrain distribution model, and obtain a plurality of first terrain features;

[0024] Analyzing the installable terrain in the terrain type of the target detection area based on N hydrological measurement indicators to obtain multiple optimal installation terrains, and searching for standard features of each optimal installation terrain in the target detection area based on a big data network to obtain multiple second terrain features; wherein the optimal installation terrain includes a watershed section, a bank side slope, and a watershed;

[0025] Calculating the Euclidean distance between each of the first terrain features and each of the second terrain features, determining the similarity between each of the first terrain features and each of the second terrain features according to the Euclidean distance, comparing the similarity between each of the first terrain features and each of the second terrain features with a preset similarity one by one, and marking the location of the first terrain features corresponding to the similarity greater than the preset similarity in the terrain distribution model, to obtain a type of planning area;

[0026] If the similarity between the second terrain feature and all the first terrain features is less than the preset similarity, the maximum similarity is extracted according to the similarity between the second terrain feature and all the first terrain features, and the location of the first terrain feature corresponding to the maximum similarity is marked to obtain the second type of planning area;

[0027] The first type of planned area and the second type of planned area are combined to obtain a first installation area.

[0028] Furthermore, in a preferred embodiment of the present invention, the construction of a hydrological detection equipment simulation model is carried out, and the hydrological detection equipment simulation model is imported into the first installation area for simulation measurement to obtain simulation signal transmission frequency band results and signal transmission interruption results, and the installation area is screened according to the simulation signal transmission frequency band results and signal transmission interruption results to obtain the second installation area, which specifically includes the following steps:

[0029] Use the Internet of Things technology to control the structured light scanning technology to perform overall structured light scanning on the hydrological detection equipment to obtain point cloud data of multiple hydrological detection equipment;

[0030] A Poisson reconstruction algorithm is introduced to perform three-dimensional reconstruction on the plurality of point cloud data, and the point cloud data is discretized into voxel three-dimensional grid data. The gradient and Laplace operator of each internal point of the grid are calculated according to the voxel three-dimensional grid data, and an equation is constructed according to the gradient and the Laplace operator to obtain the Poisson equation. Finally, the implicit function and isosurface of the Poisson equation are solved and transformed and reconstructed to obtain a simulation model of the hydrological detection equipment.

[0031] Divide the first installation area into a plurality of sub-installation areas, and sequentially import the hydrological detection equipment simulation model into each sub-installation area to perform simulated operation detection for a preset time period, and obtain simulated signal transmission frequency band results and signal transmission interruption results;

[0032] Analyze the signal transmission interruption result, extract sub-installation areas corresponding to signal transmission interruption times less than a preset number, and filter the extracted sub-installation areas based on the analog signal transmission frequency band result to obtain a plurality of filtered analog signal transmission frequency bands;

[0033] Obtaining a standard signal transmission frequency band of a hydrological detection device within a preset time period, introducing a hash algorithm to calculate a transmission error amplitude between each of the filtered analog signal transmission frequency bands and the standard signal transmission frequency band, and obtaining a plurality of hash values;

[0034] If the hash value is greater than a preset hash value, the sub-installation area corresponding to the hash value greater than the preset hash value is extracted and reorganized to obtain a second installation area.

[0035] Furthermore, in a preferred embodiment of the present invention, the calculation of the second installation area to construct a point weight descending ranking table, and at the same time establish a quadratic polynomial regression curve model based on actual water flow characteristic data and historical water flow characteristic data, analyze the quadratic polynomial regression curve model to match the final layout points in the point weight descending ranking table, and obtain the final layout plan, specifically including the following steps:

[0036] The second installation area is screened for layout points by using a particle swarm optimization algorithm, a number of particles are randomly generated in the second installation area and the initial speed and initial position of each particle are initialized, the fitness is calculated according to the local optimal position and the global optimal position of the current particle, the initial speed and the initial position are updated based on the fitness, and it is continuously iterated until a preset number of iterations is reached to obtain M installation points;

[0037] The weights of the M installation points are calculated by a weighted search algorithm to obtain M installation weight values, and the M installation weight values ​​are arranged in descending order to obtain a point weight descending arrangement table;

[0038] Obtaining the environmental conditions corresponding to each historical water flow characteristic data, re-importing the hydrological detection equipment simulation model into the M installation points based on the environmental conditions to perform secondary simulation measurement for a preset time period, obtaining simulation measurement results, and analyzing the simulation measurement results to extract actual water flow characteristic data corresponding to each installation point;

[0039] A quadratic polynomial regression algorithm is introduced to perform curve fitting on several groups of actual water flow characteristic data and several groups of historical water flow characteristic data to obtain a quadratic polynomial regression curve model, and based on the quadratic polynomial regression curve model, the fluctuation changes between the actual water flow characteristic data and the historical water flow characteristic data are analyzed and the quadratic term coefficient is calculated to obtain the fluctuation factor;

[0040] Determine whether the fluctuation factor is less than a preset fluctuation factor. If so, extract the installation point corresponding to the maximum installation weight value from the point weight descending sorting table as the optimal layout point for priority output;

[0041] If it is greater, the point weight descending sorting table is traversed to re-match the best installation point until it is less than the preset fluctuation factor, and the final layout plan is generated and uploaded to the hydrological detection equipment control terminal.

[0042] A second aspect of the present invention provides a hydrological detection equipment layout system based on the Internet of Things, wherein the hydrological detection equipment layout system based on the Internet of Things includes a memory and a processor, wherein the memory stores a hydrological detection equipment layout method program based on the Internet of Things, and when the hydrological detection equipment layout method program based on the Internet of Things is executed by the processor, the following steps are implemented:

[0043] Acquire multiple frames of continuous image information of the target detection area within a preset time period, perform feature extraction on the multiple frames of continuous image information to obtain multiple continuous feature vectors, and introduce a hidden Markov model to analyze the multiple continuous feature vectors to generate a time series dynamic model of the target detection area;

[0044] Based on the time series dynamic model, several groups of historical water flow characteristic data are obtained, and the occurrence probability of a hydrological natural disaster type is calculated in a Bayesian network algorithm according to each group of the historical water flow characteristic data, and analysis and planning are performed based on the occurrence probability to obtain a first installation area;

[0045] Constructing a hydrological detection equipment simulation model, importing the hydrological detection equipment simulation model into the first installation area for simulation measurement, obtaining simulation signal transmission frequency band results and signal transmission interruption results, and screening the installation area according to the simulation signal transmission frequency band results and signal transmission interruption results to obtain a second installation area;

[0046] By calculating the second installation area, a point weight descending list is constructed, and a quadratic polynomial regression curve model is established based on the actual water flow characteristic data and the historical water flow characteristic data. The quadratic polynomial regression curve model is analyzed to match the final layout points in the point weight descending list to obtain the final layout plan.

[0047] The present invention solves the technical defects existing in the background technology, and the beneficial technical effects of the present invention are:

[0048] Acquire multiple frames of continuous image information of the target detection area within a preset time period, and introduce a hidden Markov model for analysis to generate a time series dynamic model of the target detection area; calculate the probability of occurrence of hydrological natural disaster types in the Bayesian network algorithm based on the time series dynamic model, and analyze and plan based on the probability of occurrence to obtain a first installation area; construct a hydrological detection equipment simulation model, import the hydrological detection equipment simulation model into the first installation area for simulated measurement, and filter the installation area according to the simulated signal transmission frequency band result and the signal transmission interruption result to obtain a second installation area; establish a quadratic polynomial regression curve model based on actual water flow characteristic data and historical water flow characteristic data, analyze the quadratic polynomial regression curve model to match the final layout point in the second installation area, and obtain the final layout plan. The present invention can screen and plan the measurement layout points of the hydrological detection equipment, so that the hydrological detection equipment can achieve accurate and efficient measurement effects on the target hydrological area with high reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0050] Figure 1 A first method flow chart of a method for arranging hydrological detection equipment based on the Internet of Things is shown;

[0051] Figure 2 A second method flow chart of a method for arranging hydrological detection equipment based on the Internet of Things is shown;

[0052] Figure 3 A third method flow chart of a method for arranging hydrological detection equipment based on the Internet of Things is shown;

[0053] Figure 4 A system framework diagram of a hydrological detection equipment layout system based on the Internet of Things is shown. DETAILED DESCRIPTION

[0054] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0055] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0056] The first aspect of the present invention provides a method for arranging hydrological detection equipment based on the Internet of Things, such as Figure 1 As shown, the following steps are included:

[0057] S102: Acquire multiple frames of continuous image information of the target detection area within a preset time period, perform feature extraction on the multiple frames of continuous image information to obtain multiple continuous feature vectors, and introduce a hidden Markov model to analyze the multiple continuous feature vectors to generate a time series dynamic model of the target detection area;

[0058] S104: acquiring several groups of historical water flow characteristic data based on the time series dynamic model, calculating the occurrence probability of a hydrological natural disaster type in a Bayesian network algorithm according to each group of the historical water flow characteristic data, and performing analysis and planning based on the occurrence probability to obtain a first installation area;

[0059] S106: constructing a hydrological detection equipment simulation model, importing the hydrological detection equipment simulation model into the first installation area to perform simulation measurement, obtaining a simulation signal transmission frequency band result and a signal transmission interruption result, and screening the installation area according to the simulation signal transmission frequency band result and the signal transmission interruption result to obtain a second installation area;

[0060] S108: Construct a point weight descending ranking table by calculating the second installation area, and establish a quadratic polynomial regression curve model based on actual water flow characteristic data and historical water flow characteristic data, analyze the quadratic polynomial regression curve model to match the final layout points in the point weight descending ranking table, and obtain the final layout plan.

[0061] It should be noted that the hydrological detection equipment includes monitoring equipment such as water level meters, flow meters, rain gauges and water quality monitors, which can collect and measure data on various hydrological indicators required in the target detection area, realize diversified hydrological detection functions, and ensure hydrological measurement effects.

[0062] Furthermore, in a preferred embodiment of the present invention, the method of acquiring multiple frames of continuous image information of a target detection area within a preset time period, performing feature extraction on the multiple frames of the continuous image information to obtain multiple continuous feature vectors, and introducing a hidden Markov model to analyze the multiple continuous feature vectors to generate a time series dynamic model of the target detection area specifically includes the following steps:

[0063] The high-definition camera on the drone is controlled by the Internet of Things technology to continuously capture and shoot the target detection area, and obtain multiple frames of continuous image information of the target detection area within a preset time period;

[0064] Based on the HOG algorithm, feature extraction is performed on the continuous image information of multiple frames, each frame of image information is divided into a number of area blocks, a Sobel filter is added to the HOG algorithm to calculate the gradient size and gradient direction of the pixel points in each area block, a gradient histogram is constructed according to the gradient size and the gradient direction, and the gradient histograms constructed by the multiple frames of image information are connected to obtain multiple continuous feature vectors;

[0065] Introducing a hidden Markov model to analyze the plurality of continuous feature vectors, determining a target hidden state and a target observed variable based on each feature vector, constructing hidden states based on the target hidden state and the target observed variable, respectively, to obtain a plurality of dynamic feature sequences;

[0066] The hidden state sequence of each dynamic feature sequence is analyzed and calculated based on the Baum-Welch algorithm to obtain dynamic model parameters, and a temporal dynamic model of the target detection area is generated based on the dynamic model parameters.

[0067] It should be noted that due to the rugged geographical terrain of some detection areas, it is difficult for people to go deep into these detection areas to test the layout points of hydrological detection equipment by hiking or climbing, which may easily cause personal injury, and it is difficult to make adjustments according to the actual situation of the detection area, which reduces the measurement quality and accuracy. Therefore, continuous images can be taken by drones to construct a time series dynamic model of the target detection area. The time series dynamic model can accurately and intuitively display the water flow conditions, geographical terrain conditions and weather conditions of the target detection area, and provide real data for the layout and installation of hydrological detection equipment in a dynamic way, further ensure the accuracy of equipment installation, and improve equipment monitoring and measurement performance; continuous image features can accurately express the timing of the dynamic model, and the hidden Markov model is a statistical model that can model time series data. Combined with the Baum-Welch algorithm, it can calculate the hidden state sequence required to build a dynamic model, so that a high-quality time series dynamic model can be accurately and reliably constructed, which greatly improves the quality of virtual layout and installation point selection of hydrological detection equipment, reduces the steps and dangers of manual field point selection, and has high reliability.

[0068] Furthermore, in a preferred embodiment of the present invention, the method of obtaining several groups of historical water flow characteristic data based on the time series dynamic model, calculating the occurrence probability of a hydrological natural disaster type in a Bayesian network algorithm according to each group of the historical water flow characteristic data, and performing analysis and planning based on the occurrence probability to obtain a first installation area specifically includes the following steps:

[0069] Acquire geographical terrain information of the target detection area, and search for hydrological and natural disasters in a big data network based on the geographical terrain information to obtain a plurality of associated hydrological and natural disaster types;

[0070] Acquire several groups of historical water flow characteristic data based on the time series dynamic model of the target detection area; wherein the water flow characteristics include flow velocity, flow direction, flow rate and increase;

[0071] Constructing variable nodes and variable dependency relationships in a Bayesian network based on each group of the historical water flow characteristics and a plurality of the hydrological natural disaster types, assigning a probability to each variable node according to the variable dependency relationship, and obtaining a probability distribution of each variable node;

[0072] The state transition probability between each set of historical water flow characteristic data and each associated hydrological natural disaster type is determined through the probability distribution of each variable node, and finally the occurrence probability corresponding to each hydrological natural disaster type is determined according to the state transition probability;

[0073] Extract all types of hydrological natural disasters with a probability of occurrence greater than a preset probability of occurrence, generate search tags according to all types of hydrological natural disasters with a probability of occurrence greater than a preset probability of occurrence, and search in a big data network based on the search tags to obtain N hydrological measurement indicators;

[0074] A terrain distribution model is obtained according to a time series dynamic model of the target detection area, and the terrain distribution model is analyzed and planned based on N hydrological measurement indicators to obtain a first installation area.

[0075] It should be noted that for the purpose of hydrological measurement, preventing the occurrence of hydrological natural disasters is one of the most critical reasons. By analyzing the hydrological data measured by the hydrological detection equipment, a corresponding natural disaster prevention plan can be formulated for the target detection area. However, the hydrological natural disasters that will occur in the detection areas with different geographical terrains are also different, and the corresponding probability of occurrence is also different. If it is impossible to accurately determine the type of hydrological natural disaster that is likely to occur in the target detection area, it will be difficult to obtain the hydrological indicators that need to be measured, and it will be impossible to determine the precise equipment installation point based on the hydrological indicators, so that the hydrological detection equipment cannot accurately and reliably measure the hydrological data to provide data support for the natural disaster prevention plan. The probability and harmfulness of natural disasters in the target detection area are improved; wherein the types of hydrological natural disasters include floods, droughts, rainstorms and mudslides; the hydrological measurement indicators include water level, flow rate, rainfall and water quality; firstly, the types of hydrological natural disasters that may occur are searched out according to the geographical terrain information of the target detection area, and the probability of the occurrence of hydrological natural disasters of several groups of historical water flow characteristic data is calculated by Bayesian network analysis in the time series dynamic model, and the required hydrological measurement indicators that can play an important analytical role in the formulation of prevention plans are determined according to the types of hydrological natural disasters with a high probability of occurrence, and finally points are selected in the time series dynamic model according to the required hydrological measurement indicators. The present invention can determine the hydrological data of the points that need to be measured according to the hydrological natural disasters with a high probability of occurrence in the target detection area, thereby providing accurate and powerful data support for the formulation of hydrological natural disaster prevention plans and improving the prevention capabilities of hydrological natural disasters.

[0076] Furthermore, in a preferred embodiment of the present invention, the terrain distribution model is obtained according to the time series dynamic model of the target detection area, and the terrain distribution model is analyzed and planned based on N hydrological measurement indicators to obtain the first installation area, such as Figure 2 As shown, the specific steps include:

[0077] S202: Acquire a terrain type of a target detection area, separate a sub-model from a time series dynamic model of the target detection area, obtain a terrain distribution model, extract all terrain features based on the terrain distribution model, and obtain a plurality of first terrain features;

[0078] S204: Analyzing the installable terrain in the terrain type of the target detection area based on N hydrological measurement indicators to obtain multiple optimal installation terrains, and searching for standard features of each optimal installation terrain in the target detection area based on a big data network to obtain multiple second terrain features; wherein the optimal installation terrain includes a watershed section, a bank side slope, and a watershed;

[0079] S206: calculating the Euclidean distance between each of the first terrain features and each of the second terrain features, determining the similarity between each of the first terrain features and each of the second terrain features according to the Euclidean distance, comparing the similarity between each of the first terrain features and each of the second terrain features with a preset similarity one by one, and marking the location of the first terrain features corresponding to the similarity greater than the preset similarity in the terrain distribution model, to obtain a type of planning area;

[0080] S208: If the similarity between the second terrain feature and all the first terrain features is less than the preset similarity, extract the maximum similarity according to the similarity between the second terrain feature and all the first terrain features, and mark the location of the first terrain feature corresponding to the maximum similarity to obtain the second type of planning area;

[0081] S210: Merge the first type of planned area and the second type of planned area to obtain a first installation area.

[0082] It should be noted that in order to measure the hydrological numbers required by the hydrological measurement indicators more quickly and efficiently, there are certain terrain requirements for the installation of traditional hydrological detection equipment. For example, installing it on the watershed section can improve the accuracy of the measurement data and better play the measurement performance of the equipment. Therefore, firstly, a suitable installation terrain is analyzed for each hydrological measurement indicator, and the characteristics of the installation terrain are obtained. The characteristics of the installation terrain are matched with the required corresponding or highly similar terrain features in the terrain distribution model in the time series dynamic model, and the terrain features matched in the terrain distribution model are merged as the installable area of ​​the hydrological detection equipment, so as to achieve the effect of accurately selecting the equipment layout installation points according to the hydrological measurement indicators, and ensure that the hydrological detection equipment can synchronously obtain the required hydrological data in the target detection area according to the hydrological measurement indicators. Among them, it should be added that when all the first bottom features in the target detection area do not have a terrain corresponding to or similar to the optimal installation terrain, then the best installation terrain corresponding to the second feature in the terrain distribution model should be matched as much as possible according to the similarity. The first feature with a higher similarity makes all the points of the best installation terrain specified by the hydrological measurement indicators can meet the screening. The present invention can screen installation points in a time-series dynamic model according to hydrological measurement indicators. The screened installation points can reduce the data error rate caused by layout position deviation in traditional layout methods, further improve the accuracy of data measurement of hydrological detection equipment, and have high reliability.

[0083] Furthermore, in a preferred embodiment of the present invention, the hydrological detection equipment simulation model is constructed, and the hydrological detection equipment simulation model is imported into the first installation area for simulation measurement to obtain the simulation signal transmission frequency band result and the signal transmission interruption result, and the installation area is screened according to the simulation signal transmission frequency band result and the signal transmission interruption result to obtain the second installation area, such as Figure 3 As shown, the specific steps include:

[0084] S302: Controlling the structured light scanning technology through the Internet of Things technology to perform overall structured light scanning on the hydrological detection equipment to obtain point cloud data of multiple hydrological detection equipment;

[0085] S304: Introducing a Poisson reconstruction algorithm to perform three-dimensional reconstruction on the plurality of point cloud data, discretizing the point cloud data into voxel three-dimensional grid data, calculating the gradient and Laplace operator of each grid internal point according to the voxel three-dimensional grid data, and constructing an equation according to the gradient and the Laplace operator to obtain the Poisson equation, and finally solving the implicit function and isosurface of the Poisson equation and transforming and reconstructing them to obtain a simulation model of the hydrological detection equipment;

[0086] S306: Divide the first installation area into a plurality of sub-installation areas, and sequentially import the hydrological detection equipment simulation model into each sub-installation area to perform simulated operation detection in a preset time period, and obtain simulated signal transmission frequency band results and signal transmission interruption results;

[0087] S308: Analyze the signal transmission interruption result, extract the sub-installation area corresponding to the number of signal transmission interruptions less than the preset number, and filter the extracted sub-installation area based on the analog signal transmission frequency band result to obtain multiple filtered analog signal transmission frequency bands;

[0088] S310: obtaining a standard signal transmission frequency band of the hydrological detection equipment within a preset time period, introducing a hash algorithm to calculate the transmission error amplitude between each of the filtered analog signal transmission frequency bands and the standard signal transmission frequency band, and obtaining a plurality of hash values;

[0089] S312: If the hash value is greater than a preset hash value, extract and reorganize the sub-installation area corresponding to the hash value greater than the preset hash value to obtain a second installation area.

[0090] It should be noted that after the points of the best installation terrain have been screened, the hydrological detection equipment can be used for actual measurement in the first installation area. After solving the terrain problem, the problem of data signal transmission must also be considered. Since the existing hydrological detection equipment mainly transmits the measured data in the form of signals to fixed base stations such as hydrological monitoring points through wireless networks, and the target detection area environment may have vegetation, rocks, aquatic organisms and other obstructing factors that block signal transmission, the signal may be blocked during the signal transmission process, resulting in the inability to receive hydrological measurement data, which greatly affects the quality of hydrological measurements and the efficiency of data processing and analysis. Therefore, it is necessary to screen and exclude points with obstructions in the first installation area to solve the signal problem. Problems such as weak transmission signal and signal transmission interruption caused by transmission obstruction; among them, for the actual measurement of the hydrological detection equipment in the first installation area, if it is measured on the spot, the equipment installation will not be accurate enough, which will lead to the need to repeatedly adjust the equipment position, slow testing efficiency, time-consuming and laborious; therefore, the first installation area can be simulated and measured by constructing a simulation model of the hydrological detection equipment, so as to facilitate the repeated debugging and data recording of the operator, replacing the cumbersome steps of traditional manual field repeated debugging, with high data measurement quality and good practicality; the hash algorithm can calculate the fixed length between the data, so as to more intuitively reflect the error amplitude between the data, and the hash value is the specific representation of the error amplitude. The present invention can analyze the signal obstruction phenomenon encountered by the hydrological detection equipment in the simulation measurement process of the first installation area, and further screen out high-quality layout points according to the analysis results, so as to better ensure the signal transmission effect of the equipment and improve the accuracy of the measurement data.

[0091] Furthermore, in a preferred embodiment of the present invention, the calculation of the second installation area to construct a point weight descending ranking table, and at the same time establish a quadratic polynomial regression curve model based on actual water flow characteristic data and historical water flow characteristic data, analyze the quadratic polynomial regression curve model to match the final layout points in the point weight descending ranking table, and obtain the final layout plan, specifically including the following steps:

[0092] The second installation area is screened for layout points by using a particle swarm optimization algorithm, a number of particles are randomly generated in the second installation area and the initial speed and initial position of each particle are initialized, the fitness is calculated according to the local optimal position and the global optimal position of the current particle, the initial speed and the initial position are updated based on the fitness, and it is continuously iterated until a preset number of iterations is reached to obtain M installation points;

[0093] The weights of the M installation points are calculated by a weighted search algorithm to obtain M installation weight values, and the M installation weight values ​​are arranged in descending order to obtain a point weight descending arrangement table;

[0094] Obtaining the environmental conditions corresponding to each historical water flow characteristic data, re-importing the hydrological detection equipment simulation model into the M installation points based on the environmental conditions to perform secondary simulation measurement for a preset time period, obtaining simulation measurement results, and analyzing the simulation measurement results to extract actual water flow characteristic data corresponding to each installation point;

[0095] A quadratic polynomial regression algorithm is introduced to perform curve fitting on several groups of actual water flow characteristic data and several groups of historical water flow characteristic data to obtain a quadratic polynomial regression curve model, and based on the quadratic polynomial regression curve model, the fluctuation changes between the actual water flow characteristic data and the historical water flow characteristic data are analyzed and the quadratic term coefficient is calculated to obtain the fluctuation factor;

[0096] Determine whether the fluctuation factor is less than a preset fluctuation factor. If so, extract the installation point corresponding to the maximum installation weight value from the point weight descending sorting table as the optimal layout point for priority output;

[0097] If it is greater, the point weight descending sorting table is traversed to re-match the best installation point until it is less than the preset fluctuation factor, and the final layout plan is generated and uploaded to the hydrological detection equipment control terminal.

[0098] It should be noted that after excluding the signal shielding factors existing at each installation point in the first installation area, a second installation area is generated. At this time, the installation points can be screened in the second installation area. The particle swarm optimization algorithm can analyze all the points in the second installation area more quickly and accurately, and at the same time, a weighted search algorithm is introduced to assign a weight value reflecting the advantages and disadvantages to each installation point, providing a screening basis for the final installation of the hydrological detection equipment; then the simulation model of the hydrological detection equipment is re-imported into all the installation points in the second installation area for secondary simulation. Since the measurement results of the equipment at each installation point are different, if there is no standard reference data for further analysis of each measurement result, it is difficult to know whether there is an error in the measurement data of the point; the quadratic polynomial regression algorithm is introduced to construct a quadratic Polynomial regression curve model, since the screening of the second installation area is determined based on the time series dynamic model, and the time series dynamic model is a historical scene model that has occurred, the data of the simulation test should remain stable with the historical data and should not produce excessive fluctuations; in the curve model, the historical water flow characteristic data is curve-fitted as standard control data, and the actual water flow characteristic data measured at each point is fitted at the same time. The quadratic term coefficient calculated by the quadratic polynomial regression curve model can express the fluctuation range between the actual data of each point and the standard control data; the fluctuation factor is the degree of fluctuation of the actual water flow characteristic data relative to the historical water flow characteristic data; finally, the installation point is selected in the descending order table of the point weight according to the fluctuation factor, so as to determine the final layout plan. The present invention can perform data comparison and analysis on the simulated measurement data of the hydrological detection equipment in the second installation area, judge that the simulation test results and the historical test results remain stable, and thus re-screen the point measurement accuracy to ensure that the measurement data is stable and does not deviate, improve the measurement performance and quality of the equipment, and have high reliability.

[0099] In addition, the method for arranging hydrological detection equipment based on the Internet of Things further includes the following steps:

[0100] Obtain historical measurement data of hydrological detection equipment under different preset measurement environment condition combinations, build a prediction framework based on a deep belief network, import the historical measurement data under the different preset measurement environment condition combinations into the prediction framework for training, and obtain a trained prediction framework;

[0101] Acquire the actual measurement environment of the hydrological detection equipment, import the actual measurement environment into the trained prediction framework, and obtain the predicted measurement data under the actual measurement environment;

[0102] Acquire actual measurement data of the current hydrological detection equipment, calculate the difference between the actual measurement data and the predicted measurement data, and obtain a deviation rate threshold;

[0103] It is determined whether the deviation rate threshold is greater than a preset deviation rate threshold. If so, the hydrological detection equipment whose deviation rate threshold is greater than the preset deviation rate threshold is marked as faulty.

[0104] It should be noted that under different measurement environments, hydrological detection equipment will cause aging and damage as the operating time increases, which will lead to malfunction of the hydrological detection equipment, reduce the measurement quality of hydrological data, and make it impossible for surveyors to discover and perform maintenance in time. By using historical measurement data under different preset measurement environment conditions to predict the ideal measurement data under the current actual environment, the measurement deviation between the ideal measurement data and the actual measurement data can be used to determine whether the hydrological detection equipment has a fault, thereby realizing the automatic fault detection function, so that surveyors can perceive equipment failures in time and perform maintenance on them, thereby reducing equipment failure rates and improving expected measurement results.

[0105] In addition, the method for arranging hydrological detection equipment based on the Internet of Things further includes the following steps:

[0106] Acquire real-time weather forecast information of the target detection area, and analyze the real-time weather forecast information to obtain the real-time rainfall scale and duration in a preset time period;

[0107] According to the real-time rainfall scale of the preset time period, a number of historical rainfall amounts are extracted from the meteorological records, and the total rainfall amount within the duration is estimated by combining the number of historical rainfall amounts with the duration; wherein the rainfall scale includes showers, thunderstorms, rainstorms and typhoons;

[0108] Calculate the rising water level based on the total rainfall during the duration to obtain an actual rising water level value, and determine whether the actual rising water level value is greater than a preset rising water level value. If so, it indicates that the water level overflows and causes a flood disaster, and obtains an early warning level of the rain scale;

[0109] Retrieve preventive measures in the big data network based on the warning level of the rain scale, obtain multiple related preventive measures, and obtain the protection strength of each preventive measures plan, extract the preventive measures plan whose protection strength is greater than the preset protection strength, and obtain the preventive measures plan after a screening;

[0110] The cosine similarity is introduced to calculate the applicability between the preventive measures scheme after the primary screening and the terrain of the target detection area, and multiple applicability rates are obtained. If the applicability rate is greater than the preset applicability rate, the corresponding preventive measures scheme greater than the preset applicability rate is extracted to obtain the preventive measures scheme after the secondary screening;

[0111] The preventive measures scheme corresponding to the maximum protection intensity is extracted from the preventive measures scheme after the secondary screening and output as the final preventive measures of the real-time rain scale.

[0112] It should be noted that there are many measurement areas where hydrological detection equipment is applicable, such as rivers, lakes and shallow seas. Due to the influence of the terrain, the water level in each area will rise on rainy days of different scales, thus causing floods and endangering people's lives. Since the rising water level that causes floods in each area is closely related to the scale of rain, it is necessary to determine the warning level of the current rain scale through the actual rising water level value, and then match the associated preventive measures according to the warning level, and consider the protection strength of the preventive measures and the applicability of the target detection area for gradual screening, until the most matching and best preventive measures are selected as a powerful means to resist the occurrence of floods, achieve high-quality flood control effects, further reduce the losses and harm caused by floods, and improve the flood control safety factor.

[0113] A second aspect of the present invention provides a hydrological detection equipment layout system based on the Internet of Things, wherein the hydrological detection equipment layout system based on the Internet of Things comprises a memory 41 and a processor 42, wherein the memory 41 stores a hydrological detection equipment layout method program based on the Internet of Things, and when the hydrological detection equipment layout method program based on the Internet of Things is executed by the processor 42, Figure 4 As shown, implement the following steps:

[0114] Acquire multiple frames of continuous image information of the target detection area within a preset time period, perform feature extraction on the multiple frames of continuous image information to obtain multiple continuous feature vectors, and introduce a hidden Markov model to analyze the multiple continuous feature vectors to generate a time series dynamic model of the target detection area;

[0115] Based on the time series dynamic model, several groups of historical water flow characteristic data are obtained, and the occurrence probability of a hydrological natural disaster type is calculated in a Bayesian network algorithm according to each group of the historical water flow characteristic data, and analysis and planning are performed based on the occurrence probability to obtain a first installation area;

[0116] Constructing a hydrological detection equipment simulation model, importing the hydrological detection equipment simulation model into the first installation area for simulation measurement, obtaining simulation signal transmission frequency band results and signal transmission interruption results, and screening the installation area according to the simulation signal transmission frequency band results and signal transmission interruption results to obtain a second installation area;

[0117] By calculating the second installation area, a point weight descending list is constructed, and a quadratic polynomial regression curve model is established based on the actual water flow characteristic data and the historical water flow characteristic data. The quadratic polynomial regression curve model is analyzed to match the final layout points in the point weight descending list to obtain the final layout plan.

[0118] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for layout of hydrological detection equipment based on the Internet of Things, characterized in that: The following steps are involved: Acquire multiple frames of continuous image information of the target detection area within a preset time period, perform feature extraction on the multiple frames of continuous image information to obtain multiple continuous feature vectors, and introduce a hidden Markov model to analyze the multiple continuous feature vectors to generate a time series dynamic model of the target detection area; Based on the time series dynamic model, several groups of historical water flow characteristic data are obtained, and the occurrence probability of a hydrological natural disaster type is calculated in a Bayesian network algorithm according to each group of the historical water flow characteristic data, and analysis and planning are performed based on the occurrence probability to obtain a first installation area; Constructing a hydrological detection equipment simulation model, importing the hydrological detection equipment simulation model into the first installation area for simulation measurement, obtaining simulation signal transmission frequency band results and signal transmission interruption results, and screening the installation area according to the simulation signal transmission frequency band results and signal transmission interruption results to obtain a second installation area; A point weight descending ranking table is constructed by calculating the second installation area, and a quadratic polynomial regression curve model is established based on actual water flow characteristic data and historical water flow characteristic data, and the quadratic polynomial regression curve model is analyzed to match the final layout points in the point weight descending ranking table to obtain a final layout plan; The method of obtaining multiple frames of continuous image information of the target detection area within a preset time period and performing feature extraction on the multiple frames of continuous image information to obtain multiple continuous feature vectors specifically includes the following steps: The high-definition camera on the drone is controlled by the Internet of Things technology to continuously capture and shoot the target detection area, and obtain multiple frames of continuous image information of the target detection area within a preset time period; Based on the HOG algorithm, feature extraction is performed on the continuous image information of multiple frames, each frame of image information is divided into several area blocks, a Sobel filter is added to the HOG algorithm to calculate the gradient size and gradient direction of the pixel points in each area block, a gradient histogram is constructed according to the gradient size and the gradient direction, and the gradient histograms constructed by the multiple frames of image information are connected to obtain multiple continuous feature vectors.

2. According to the method for layout of hydrological detection equipment based on the Internet of Things in claim 1, it is characterized in that: The hidden Markov model is introduced to analyze multiple continuous feature vectors to generate a time series dynamic model of the target detection area, which specifically includes the following steps: Introducing a hidden Markov model to analyze the plurality of continuous feature vectors, determining a target hidden state and a target observed variable based on each feature vector, constructing hidden states based on the target hidden state and the target observed variable, respectively, to obtain a plurality of dynamic feature sequences; The hidden state sequence of each dynamic feature sequence is analyzed and calculated based on the Baum-Welch algorithm to obtain dynamic model parameters, and a temporal dynamic model of the target detection area is generated based on the dynamic model parameters.

3. The method for arranging hydrological detection equipment based on the Internet of Things according to claim 1 is characterized in that: The method of acquiring several groups of historical water flow characteristic data based on the time series dynamic model, calculating the occurrence probability of a hydrological natural disaster type in a Bayesian network algorithm according to each group of the historical water flow characteristic data, and performing analysis and planning based on the occurrence probability to obtain a first installation area specifically includes the following steps: Acquire geographical terrain information of the target detection area, and search for hydrological and natural disasters in a big data network based on the geographical terrain information to obtain a plurality of associated hydrological and natural disaster types; Acquire several groups of historical water flow characteristic data based on the time series dynamic model of the target detection area; wherein the water flow characteristics include flow velocity, flow direction, flow rate and increase; Constructing variable nodes and variable dependency relationships in a Bayesian network based on each group of the historical water flow characteristics and a plurality of the hydrological natural disaster types, assigning a probability to each variable node according to the variable dependency relationship, and obtaining a probability distribution of each variable node; The state transition probability between each set of historical water flow characteristic data and each associated hydrological natural disaster type is determined through the probability distribution of each variable node, and finally the occurrence probability corresponding to each hydrological natural disaster type is determined according to the state transition probability; Extract all types of hydrological natural disasters with a probability of occurrence greater than a preset probability of occurrence, generate search tags according to all types of hydrological natural disasters with a probability of occurrence greater than a preset probability of occurrence, and search in a big data network based on the search tags to obtain N hydrological measurement indicators; A terrain distribution model is obtained according to a time series dynamic model of the target detection area, and the terrain distribution model is analyzed and planned based on N hydrological measurement indicators to obtain a first installation area.

4. The method for arranging hydrological detection equipment based on the Internet of Things according to claim 3 is characterized in that: The method of acquiring a terrain distribution model according to a time series dynamic model of the target detection area, analyzing and planning the terrain distribution model based on N hydrological measurement indicators, and obtaining a first installation area specifically includes the following steps: Acquire a terrain type of a target detection area, separate a sub-model from a time series dynamic model of the target detection area to obtain a terrain distribution model, extract all terrain features based on the terrain distribution model, and obtain a plurality of first terrain features; Analyzing the installable terrain in the terrain type of the target detection area based on N hydrological measurement indicators to obtain multiple optimal installation terrains, and searching for standard features of each optimal installation terrain in the target detection area based on a big data network to obtain multiple second terrain features; wherein the optimal installation terrain includes a watershed section, a bank side slope, and a watershed; Calculating the Euclidean distance between each of the first terrain features and each of the second terrain features, determining the similarity between each of the first terrain features and each of the second terrain features according to the Euclidean distance, comparing the similarity between each of the first terrain features and each of the second terrain features with a preset similarity one by one, and marking the location of the first terrain features corresponding to the similarity greater than the preset similarity in the terrain distribution model, to obtain a type of planning area; If the similarity between the second terrain feature and all the first terrain features is less than the preset similarity, the maximum similarity is extracted according to the similarity between the second terrain feature and all the first terrain features, and the location of the first terrain feature corresponding to the maximum similarity is marked to obtain the second type of planning area; The first type of planned area and the second type of planned area are combined to obtain a first installation area.

5. The method for arranging hydrological detection equipment based on the Internet of Things according to claim 1 is characterized in that: The construction of the hydrological detection equipment simulation model, importing the hydrological detection equipment simulation model into the first installation area for simulation measurement, obtaining the simulated signal transmission frequency band result and the signal transmission interruption result, and screening the installation area according to the simulated signal transmission frequency band result and the signal transmission interruption result to obtain the second installation area, specifically includes the following steps: Use the Internet of Things technology to control the structured light scanning technology to perform overall structured light scanning on the hydrological detection equipment to obtain point cloud data of multiple hydrological detection equipment; A Poisson reconstruction algorithm is introduced to perform three-dimensional reconstruction on the plurality of point cloud data, and the point cloud data is discretized into voxel three-dimensional grid data. The gradient and Laplace operator of each internal point of the grid are calculated according to the voxel three-dimensional grid data, and an equation is constructed according to the gradient and the Laplace operator to obtain the Poisson equation. Finally, the implicit function and isosurface of the Poisson equation are solved and transformed and reconstructed to obtain a simulation model of the hydrological detection equipment. Divide the first installation area into a plurality of sub-installation areas, and import the hydrological detection equipment simulation model into each sub-installation area in turn to perform simulated operation detection in a preset time period, and obtain simulated signal transmission frequency band results and signal transmission interruption results; Analyze the signal transmission interruption result, extract sub-installation areas corresponding to signal transmission interruption times less than a preset number, and filter the extracted sub-installation areas based on the analog signal transmission frequency band result to obtain a plurality of filtered analog signal transmission frequency bands; Obtaining a standard signal transmission frequency band of a hydrological detection device within a preset time period, introducing a hash algorithm to calculate a transmission error amplitude between each of the filtered analog signal transmission frequency bands and the standard signal transmission frequency band, and obtaining a plurality of hash values; If the hash value is greater than a preset hash value, the sub-installation area corresponding to the hash value greater than the preset hash value is extracted and reorganized to obtain a second installation area.

6. The method for arranging hydrological detection equipment based on the Internet of Things according to claim 1 is characterized in that: The method of constructing a point weight descending arrangement table by calculating the second installation area, establishing a quadratic polynomial regression curve model based on actual water flow characteristic data and historical water flow characteristic data, and analyzing the quadratic polynomial regression curve model to match the final layout points in the point weight descending arrangement table to obtain the final layout plan specifically includes the following steps: The second installation area is screened for layout points by using a particle swarm optimization algorithm, a number of particles are randomly generated in the second installation area and the initial speed and initial position of each particle are initialized, the fitness is calculated according to the local optimal position and the global optimal position of the current particle, the initial speed and the initial position are updated based on the fitness, and it is continuously iterated until a preset number of iterations is reached to obtain M installation points; The weights of the M installation points are calculated by a weighted search algorithm to obtain M installation weight values, and the M installation weight values ​​are arranged in descending order to obtain a point weight descending arrangement table; Obtaining the environmental conditions corresponding to each historical water flow characteristic data, re-importing the hydrological detection equipment simulation model into the M installation points based on the environmental conditions to perform secondary simulation measurement for a preset time period, obtaining simulation measurement results, and analyzing the simulation measurement results to extract actual water flow characteristic data corresponding to each installation point; A quadratic polynomial regression algorithm is introduced to perform curve fitting on several groups of actual water flow characteristic data and several groups of historical water flow characteristic data to obtain a quadratic polynomial regression curve model, and based on the quadratic polynomial regression curve model, the fluctuation changes between the actual water flow characteristic data and the historical water flow characteristic data are analyzed and the quadratic term coefficient is calculated to obtain the fluctuation factor; Determine whether the fluctuation factor is less than a preset fluctuation factor. If so, extract the installation point corresponding to the maximum installation weight value from the point weight descending sorting table as the optimal layout point for priority output; If it is greater, the point weight descending sorting table is traversed to re-match the best installation point until it is less than the preset fluctuation factor, and the final layout plan is generated and uploaded to the hydrological detection equipment control terminal.

7. A hydrological detection equipment layout system based on the Internet of Things, characterized in that: The hydrological detection equipment layout system based on the Internet of Things includes a memory and a processor, wherein the memory stores a hydrological detection equipment layout method program based on the Internet of Things, and when the hydrological detection equipment layout method program based on the Internet of Things is executed by the processor, the following steps are implemented: Acquire multiple frames of continuous image information of the target detection area within a preset time period, perform feature extraction on the multiple frames of continuous image information to obtain multiple continuous feature vectors, and introduce a hidden Markov model to analyze the multiple continuous feature vectors to generate a time series dynamic model of the target detection area; Based on the time series dynamic model, several groups of historical water flow characteristic data are obtained, and the occurrence probability of a hydrological natural disaster type is calculated in a Bayesian network algorithm according to each group of the historical water flow characteristic data, and analysis and planning are performed based on the occurrence probability to obtain a first installation area; Constructing a hydrological detection equipment simulation model, importing the hydrological detection equipment simulation model into the first installation area for simulation measurement, obtaining simulation signal transmission frequency band results and signal transmission interruption results, and screening the installation area according to the simulation signal transmission frequency band results and signal transmission interruption results to obtain a second installation area; A point weight descending ranking table is constructed by calculating the second installation area, and a quadratic polynomial regression curve model is established based on actual water flow characteristic data and historical water flow characteristic data, and the quadratic polynomial regression curve model is analyzed to match the final layout points in the point weight descending ranking table to obtain a final layout plan; The method of obtaining multiple frames of continuous image information of the target detection area within a preset time period and performing feature extraction on the multiple frames of continuous image information to obtain multiple continuous feature vectors specifically includes the following steps: The high-definition camera on the drone is controlled by the Internet of Things technology to continuously capture and shoot the target detection area, and obtain multiple frames of continuous image information of the target detection area within a preset time period; Based on the HOG algorithm, feature extraction is performed on the continuous image information of multiple frames, each frame of image information is divided into several area blocks, a Sobel filter is added to the HOG algorithm to calculate the gradient size and gradient direction of the pixel points in each area block, a gradient histogram is constructed according to the gradient size and the gradient direction, and the gradient histograms constructed by the multiple frames of image information are connected to obtain multiple continuous feature vectors.

Images