CN117238173A - A comprehensive method and system for dynamic analysis of spatio-temporal hot spots in ship accidents - Google Patents

Abstract

The invention discloses a comprehensive method and system for dynamic analysis of ship accident spatio-temporal hot spots, which relates to the technical field of maritime traffic safety and includes: receiving ship accident data, performing data filtering and data conversion on the ship accident data, and obtaining ship accident processing data. Among them, the ship accident data includes the ship type in the ship accident, event summary, gross tonnage, classification society, accident coordinates, accident time, as well as the number of casualties, and traffic flow data of the area where the ship accident is located; based on the ship accident processing data , locate the ship accident, and obtain the geographical distribution of the ship accident. Based on the geographical distribution of the ship accident, density analysis, spatial autocorrelation analysis and three-dimensional spatio-temporal hotspot analysis are performed on the ship accident; density analysis, spatial autocorrelation analysis and Three-dimensional spatio-temporal hotspot analysis was used to adjust parameters, and after multiple screenings, the spatio-temporal hotspot output results of ship accidents were obtained.

Description

A comprehensive method and system for dynamic analysis of spatio-temporal hot spots in ship accidents

Technical field

The invention relates to the technical field of maritime traffic safety, specifically a comprehensive method and system for dynamic analysis of spatio-temporal hot spots in ship accidents.

Background technique

With the rapid development of economic globalization, shipping plays an important role in world trade. About 80% of global cargo trade is completed by international shipping. Since shipping activities are carried out in a complex and dangerous water environment, which can easily lead to ship accidents and may cause serious economic losses, casualties and marine pollution, ship safety has always been the focus of the international shipping community. Studying the spatio-temporal hotspot distribution and evolution of ship accidents will help the maritime department intuitively understand the traffic safety status of ships within their jurisdiction, and is of great significance for ensuring the navigation safety of ships, especially in early warning and forecasting.

Currently, the research methods for the distribution of ship accident hot spots mainly include traditional statistical analysis and spatial analysis based on geographic information systems. The former determines the high-incidence areas of accidents through statistical analysis based on the geographical coordinates of the accident, while the latter uses geographic information system technology to visually present the distribution characteristics of accident hot spots on the map. Most of these studies conduct macro-analysis of accidents over a long span of time to explore the two-dimensional spatial distribution characteristics of accidents. However, few researchers pay attention to the three-dimensional spatio-temporal hot spot distribution of ship accidents and the evolution of accident hot spots over time.

Contents of the invention

In order to solve the deficiencies mentioned in the above background technology, the purpose of the present invention is to provide a comprehensive dynamic analysis method and system for spatio-temporal hot spots of ship accidents, based on the historical big data of ship accidents, from the accident density, aggregation degree, and spatio-temporal hot spots. Starting from the perspective of trend and density, density analysis, spatial autocorrelation analysis and three-dimensional spatio-temporal hot spot analysis methods are used to more comprehensively and scientifically identify the spatio-temporal distribution and evolution characteristics of ship accident hot spots.

The purpose of the present invention can be achieved through the following technical solutions: a comprehensive dynamic analysis method of spatio-temporal hot spots in ship accidents, the method includes the following steps:

Receive ship accident data, perform data filtering and data conversion on the ship accident data, and obtain ship accident processing data, wherein the ship accident data includes ship type, event summary, gross tonnage, classification society, accident coordinates, The time of the accident, the number of casualties, and traffic flow data in the area where the ship accident occurred;

Based on the ship accident processing data, the ship accident is located and the geographical distribution of the ship accident is obtained. Based on the geographical distribution of the ship accident, density analysis, spatial autocorrelation analysis and three-dimensional spatio-temporal hotspot analysis are performed on the ship accident;

The parameters of density analysis, spatial autocorrelation analysis and three-dimensional spatio-temporal hot spot analysis were adjusted respectively. After multiple screenings, the output results of ship accident spatio-temporal hot spots were obtained.

Preferably, the data filtering includes filtering out accident data without coordinates in the ship accident data, clearing abnormal data, and correcting erroneous data. The data is converted into ship accident longitude and latitude coordinate data in a degree format and can be converted into geographic information system software. Recognized latitude and longitude coordinate data in decimal format.

Preferably, the density analysis includes; kernel density estimation and point density analysis, used to determine the density level and risk boundary of ship accidents.

Preferably, the calculation formula of the kernel density estimate is:

Among them, f (x, y) is the estimated density at the (x, y) position, s is the number of positions, h is the bandwidth, K is the kernel function, and λ _i is the distance from the position (x, y) to the i-th position. distance;

The calculation formula of point density analysis F is:

Among them, ρ is the neighborhood radius, N(ρ) is the number of ship accidents in a circle with the center of unit k as the center and the neighborhood radius ρ, and M is the ship traffic density.

Preferably, the spatial autocorrelation analysis includes global spatial autocorrelation analysis and local spatial autocorrelation analysis.

Preferably, the calculation formula of the global spatial autocorrelation analysis I is:

Among them, n is the total number of accidents in the study area, x _i and x _j represent the i-th and j-th spatial observation values respectively, w _ij represents the elements of the spatial binary adjacency matrix, the weight is the neighborhood relationship existing between position i and its adjacent position j, W represents the sum of all accidents in w _ij ,/>

Local spatial autocorrelation analysis The calculation formula is:

Among them, S is the standard deviation of x _j , Local spatial autocorrelation analysis classifies the degree of accident clustering into seven categories, hot or cold spots with 90%, 95% and 99% confidence levels, and points with no significant clustering.

Preferably, the three-dimensional spatio-temporal hot spot analysis includes: emerging hot spot analysis, spatio-temporal hot spot trend analysis and spatio-temporal kernel density estimation analysis.

Preferably, the emerging hotspot analysis can explain and classify accident hotspots in more detail, and can also present changes in different types of hotspots. The emerging hotspot analysis divides the results into new, continuous, enhanced, permanent, Reduce, disperse, oscillate and historical hot and cold spots. The spatio-temporal hot spot trend analysis can identify statistically significant accident hot spot changing trends. The spatio-temporal hot spot trend analysis increases the z-score of hot spots or cold spots over time. or decreasing trends are divided into seven categories, increasing or decreasing trends with 90%, 95% and 99% confidence levels, and no significant trend;

The calculation formula of spatiotemporal kernel density estimation is:

in, is the density estimate at position (x, y, t), the x and y dimensions are space, the t dimension is time, ds is the spatial bandwidth, dt is the time bandwidth, d _s ² d is the density estimate/> The density value obtained by multiplying by n is expressed as the number of events per unit of space and time. The calculation formulas of the kernel functions Ls and Lt are:

Among them, u is the difference rate between longitude xi and xi+1, v is the difference rate between latitude yi and yi+1, and p is the difference rate between time ti and ti+1.

Preferably, the process of obtaining the output results of ship accident spatio-temporal hot spots:

In the parameter adjustment process of kernel density estimation, the results of kernel density estimation are jointly affected by the kernel function K, bandwidth h, and unit size. The bandwidth affects the smoothness of the density surface, while the unit size affects the roughness of the generated surface. It needs to be After multiple screenings, a group was finally obtained that best matched the output results of ship accident nuclear density estimation;

During the parameter adjustment process of point density analysis, the results of point density analysis are jointly affected by the neighborhood radius ρ and ship traffic density M, and need to be screened multiple times to finally obtain a group that best matches the output results of ship accident point density analysis;

In the process of creating a space-time cube, emerging space-time hotspot analysis, space-time hotspot trend analysis, and parameter adjustment of space-time kernel density estimation, the analysis results are affected by the time and space size of the space-time box, and need to be screened multiple times to finally obtain the most suitable A set of output results of spatio-temporal hotspot analysis of ship accidents;

The output results that are consistent with the ship accident nuclear density estimation, the output results that are most consistent with the ship accident point density analysis, and the output results that are most consistent with the ship accident spatio-temporal hotspot analysis are combined to obtain the ship accident spatio-temporal hotspot output results.

In the second aspect, in order to achieve the above purpose, the present invention discloses a comprehensive spatio-temporal hot spot dynamic analysis system for ship accidents, including:

Data processing module: used to receive ship accident data, perform data filtering and data conversion on the ship accident data, and obtain ship accident processing data. The ship accident data includes the ship type, event summary, total tonnage, and ship accident data. Class society, accident coordinates, accident time, number of casualties, and traffic flow data in the area where the ship accident occurred;

Accident analysis module: used to locate ship accidents based on ship accident processing data and obtain the geographical distribution of ship accidents. Based on the geographical distribution of ship accidents, conduct density analysis, spatial autocorrelation analysis and three-dimensional spatio-temporal hotspot analysis of ship accidents. ;

Result output module: used to adjust parameters for density analysis, spatial autocorrelation analysis and three-dimensional spatiotemporal hotspot analysis respectively. After multiple screenings, the spatiotemporal hotspot output results of ship accidents are obtained.

Beneficial effects of the present invention:

This invention can realize macro- and micro-analysis of the spatial distribution of ship accidents, can simultaneously carry out spatial hotspot analysis and three-dimensional dynamic spatio-temporal hotspot analysis of accidents, can intuitively and accurately present the spatiotemporal pattern of ship accidents, and identify several accident-prone areas. , as well as the evolution of hotspot distribution over time, thus providing a basic reference for maritime safety departments, relevant governments and policymakers to establish reasonable and effective ship accident management strategies, and helping to take more targeted accident prevention measures.

Description of drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings needed to describe the embodiments or the prior art. Obviously, for those of ordinary skill in the art, Speaking of which, other drawings can be obtained based on these drawings without any creative effort;

Figure 1 is a schematic flow diagram of the method of the present invention;

Figure 2 is a schematic flow diagram of a comprehensive spatio-temporal hot spot dynamic analysis method for ship accidents according to the present invention;

Figure 3 is a schematic structural diagram of the space-time cube of the present invention;

Figure 4 is a conversion method and example diagram of accident longitude and latitude coordinate data in the authoritative ship accident database of the present invention;

Figure 5 is a schematic structural diagram of the system of the present invention;

Figure 6 is a global autocorrelation analysis result diagram of ship accidents according to the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

As shown in Figure 1, a comprehensive dynamic analysis method of spatio-temporal hot spots in ship accidents includes the following steps:

Receive ship accident data, perform data filtering and data conversion on the ship accident data, and obtain ship accident processing data, wherein the ship accident data includes ship type, event summary, gross tonnage, classification society, accident coordinates, The time of the accident, the number of casualties, and traffic flow data in the area where the ship accident occurred;

Based on the ship accident processing data, the ship accident is located and the geographical distribution of the ship accident is obtained. Based on the geographical distribution of the ship accident, density analysis, spatial autocorrelation analysis and three-dimensional spatio-temporal hotspot analysis are performed on the ship accident;

The parameters of density analysis, spatial autocorrelation analysis and three-dimensional spatio-temporal hot spot analysis were adjusted respectively. After multiple screenings, the output results of ship accident spatio-temporal hot spots were obtained.

It should be further explained that during the specific implementation process: the above method specifically includes:

Step 1. Original data collection.

1.1. The original data comes from the authoritative ship accident database, including but not limited to the type of ship in the ship accident, accident consequences, event summary, gross tonnage, classification society, accident coordinates, accident time, number of casualties, and location of the ship accident. Regional traffic flow data, etc. Reliable data sources are the basis for studying ship accidents. Ship accident databases can be divided into public databases managed by the International Maritime Organization, commercial databases managed by classification societies, and national databases managed by government departments and agencies. Ship accidents collected from the above databases The data is more comprehensive and accurate.

Step 2. Data filtering and transformation.

2.1. Data filtering, including filtering out accident data without coordinates in ship accident data, clearing abnormal data, and correcting erroneous data.

2.2. Data conversion: Convert the longitude and latitude coordinate data of the ship accident in degree and minute format (such as 140°57.20’) into longitude and latitude coordinate data in decimal format (such as 140.96) that can be recognized by the geographic information system software. Geographic information system analysis software can only recognize longitude and latitude coordinate data in decimal form, and the data format of longitude and latitude coordinates of ship accidents is usually degrees and minutes. The filtered ship accident data is converted into coordinate data in decimal form.

Step 3. Based on the geographical information system, establish a spatio-temporal hotspot dynamic analysis method.

3.1. Based on the geographical information system, establish a dynamic analysis method of spatio-temporal hot spots, including: (1) Locate the accident according to the longitude and latitude coordinates in the ship accident data, visualize the geographical distribution of the ship accident, and display the geographical distribution of the accident as subsequent hot spots Provide basis for analysis; (2) density analysis; (3) spatial autocorrelation analysis; (4) three-dimensional spatio-temporal hot spot analysis.

3.2. Density analysis, including; kernel density estimation and point density analysis, aims to determine the density level and risk boundary of ship accidents. The spatial analysis unit can be defined arbitrarily using kernel density estimation and point density analysis methods.

Among them, kernel density estimation is an effective accident density identification method. The hot spot refers to the area with the highest accident density level. Kernel density estimation is a non-parametric method for estimating the unknown density function of random variables. Compared with the parameter estimation method , this method does not require assumptions about the density function of the data set. Kernel density estimation is organically linked to the spatial distribution of ship accidents and the geographical environment, and can be used to determine the density level of ship accidents.

The calculation formula of kernel density estimation is:

Among them, f (x, y) is the estimated density at the (x, y) position, s is the number of positions, h is the bandwidth, K is the kernel function, and λ _i is the distance from the position (x, y) to the i-th position. distance.

The hot spot in point density analysis refers to the area with the largest number of accidents in the neighborhood. Point density analysis divides the number of points in a certain neighborhood by the neighborhood area, and takes into account the ship traffic density in the neighborhood. The calculation formula for ship accident density in a certain unit during a time period is:

Among them, ρ is the neighborhood radius, N(ρ) is the number of ship accidents in a circle with the center of unit k as the center and the neighborhood radius ρ, and M is the ship traffic density.

3.3. Spatial autocorrelation analysis is an analysis method that determines consistent object groups based on the number of object attributes. It includes: global spatial autocorrelation analysis and local spatial autocorrelation analysis, which can make up for the lack of statistical significance in the density analysis results.

Among them, the global autocorrelation analysis method is used to describe the overall distribution of a certain research object and determine whether the research object as a whole has spatial aggregation characteristics. It is carried out using global Moran's I statistical analysis. The global Moran's I method is based on the covariance of the statistical correlation coefficient. relationship, its calculation formula is:

Among them, n is the total number of accidents in the study area, x _i and x _j represent the i-th and j-th spatial observation values respectively, w _ij represents the elements of the spatial binary adjacency matrix, the weight is the neighborhood relationship existing between position i and its adjacent position j, W represents the sum of all accidents in w _ij ,/>

The local spatial autocorrelation analysis method can clarify the specific accident gathering range and identify the hot spots of accidents. Getis-Ord Gi* is a typical local spatial autocorrelation analysis technology, and its calculation formula is:

Among them, S is the standard deviation of x _j , Local spatial autocorrelation analysis divides the degree of accident clustering into seven categories, hot or cold spots with 90%, 95% and 99% confidence levels, and points with no significant clustering;

3.4. Three-dimensional spatio-temporal hot spot analysis, including: emerging hot spot analysis, spatio-temporal hot spot trend analysis and spatio-temporal kernel density estimation analysis. The purpose is to coordinate the spatial and temporal attributes of accidents, analyze the demise of old hot spots and the formation of new hot spots, and analyze accidents. Development trends and analysis of changes in accident density over time.

Among them, the emerging hot spot analysis and spatio-temporal hot spot trend analysis methods extend the Getis-Ord Gi* statistics to combine the time dimension of accident data. These two methods combine the Getis-Ord Gi* statistics and the Mann-Kendall trend test. It can not only evaluate the location and degree of accident spatial clustering, but also evaluate the changing trend of space-time box time series. The input data set of these two analysis methods is the space-time of NetCDF data format (that is, the data format used to store multi-dimensional scientific data). cube. The definition of a hotspot in emerging hotspot analysis and spatiotemporal hotspot trend analysis is that the local summary statistic of a spacetime box is significantly higher than the expected local statistic, that is, the local statistic calculated based on the assumption of complete spatial randomness, and the difference is too large. It is not the result of a random process. Emerging hotspot analysis can explain and classify accident hotspots in more detail, and can also present changes in different types of hotspots. Emerging hotspot analysis divides the results into new, continuous, enhanced, and permanent. , reduced, dispersed, oscillating and historical hot and cold spots, spatiotemporal hotspot trend analysis can identify statistically significant accident hotspot changing trends, spatiotemporal hotspot trend analysis will increase the z-score of hotspots or cold spots over time or Decreasing trends are divided into seven categories, increasing or decreasing trends with 90%, 95% and 99% confidence levels, and no significant trend.

The spatio-temporal kernel density estimation analysis breaks through the limitation that the above-mentioned hotspot analysis technology cannot study the change of ship accident density over time, and can realize the microscopic analysis of the evolution process of ship accident density over time. The calculation formula is:

in, is the density estimate at position (x, y, t), the x and y dimensions are space, the t dimension is time, ds is the spatial bandwidth, dt is the time bandwidth, d _s ² d is the density estimate/> The density value obtained by multiplying by n is expressed as the number of events per unit of space and time. The calculation formulas of the kernel functions Ls and Lt are:

Among them, u is the difference rate between longitude xi and xi+1, v is the difference rate between latitude yi and yi+1, p is the difference rate between time ti and ti+1;

Step 4. Set parameters to obtain the final spatio-temporal hot spot dynamic analysis method.

4.1. Based on the established spatio-temporal hot spot dynamic analysis method, adjust the parameters to meet the accuracy requirements and obtain the final spatio-temporal hot spot dynamic analysis method;

4.2. During the parameter adjustment process of kernel density estimation, the results of kernel density estimation are jointly affected by the kernel function K, bandwidth h, and unit size. Commonly used kernel functions include Gaussian kernel function and rectangular kernel function. Bandwidth affects the density surface. Smoothness, and the unit size affects the roughness of the generated surface, which requires multiple screenings to finally select a group that best matches the output results of the ship accident nuclear density estimation;

During the parameter adjustment process of point density analysis, the results of point density analysis are jointly affected by the neighborhood radius ρ and ship traffic density M, and need to be screened multiple times to finally select a group that best matches the output results of ship accident point density analysis; In the process of creating a space-time cube, emerging space-time hotspot analysis, space-time hotspot trend analysis, and parameter adjustment of space-time kernel density estimation, the analysis results are affected by the time and space size of the space-time box, and require multiple screenings to finally select the most suitable A set of ship accident spatio-temporal hotspot analysis output results combines the ship accident spatio-temporal hot spot output results that are consistent with the ship accident nuclear density estimation output, the most consistent ship accident point density analysis output, and the most consistent ship accident spatio-temporal hot spot analysis output results.

In this embodiment, the space-time cube (x, y, t) is a three-dimensional cube composed of space-time boxes. This tool can aggregate point input elements (ie, the coordinates and time information of ship accidents) into space-time boxes, where the x dimension The y dimension represents space and the t dimension represents time. The space-time box time series can be obtained through spatial positioning, and the space-time box time series is a vertical column. The number of data points contained in each column is the number of geographic events that occurred per unit time step. Each space-time bin has a unique location ID and holds an aggregate value, including the number of data points or other summary statistics for the specified attribute. The space-time bin The time series can visually display the changing trend of the number of ship accidents geographically over time.

On the other hand, as shown in Figure 5, embodiments of the present invention also provide a comprehensive spatio-temporal hotspot dynamic analysis system for ship accidents, including:

Data processing module: used to receive ship accident data, perform data filtering and data conversion on the ship accident data, and obtain ship accident processing data. The ship accident data includes the ship type, event summary, total tonnage, and ship accident data. Class society, accident coordinates, accident time, number of casualties, and traffic flow data in the area where the ship accident occurred;

Accident analysis module: used to locate ship accidents based on ship accident processing data and obtain the geographical distribution of ship accidents. Based on the geographical distribution of ship accidents, conduct density analysis, spatial autocorrelation analysis and three-dimensional spatio-temporal hotspot analysis of ship accidents. ;

Result output module: used to adjust parameters for density analysis, spatial autocorrelation analysis and three-dimensional spatiotemporal hotspot analysis respectively. After multiple screenings, the spatiotemporal hotspot output results of ship accidents are obtained.

Based on the same inventive concept, the present invention also provides a computer device. The computer device includes: one or more processors and a memory for storing one or more computer programs; the program includes program instructions, and the processor is used to execute the memory Stored program instructions. The processor may be a Central Processing Unit (CPU), or other general-purpose processor, Digital Signal Processor (DSP), Application Specific Integrated Circuit (ASIC), or field programmable Gate array (Field-Programmable GateArray, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, are used to implement one or more instructions, specifically It is used to load and execute one or more instructions in the computer storage medium to implement the above method.

It should be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium, the storage medium stores a computer program, and the computer program executes the above method when run by a processor. The storage medium may be any combination of one or more computer-readable media. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electrical, magnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples (non-exhaustive list) of computer readable storage media include: electrical connections having one or more conductors, portable computer disks, hard drives, random access memory (RAM), read only memory (ROM), Erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium that contains or stores a program for use by or in conjunction with an instruction execution system, apparatus, or device.

In the description of this specification, reference to the terms "one embodiment," "example," "specific example," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one aspect of the disclosure. in an embodiment or example. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

The basic principles, main features, and advantages of the present disclosure have been shown and described above. Those skilled in the industry should understand that the present disclosure is not limited by the above embodiments. What is described in the above embodiments and descriptions only illustrates the principles of the present disclosure. Without departing from the spirit and scope of the present disclosure, the present disclosure will also have other features. Various changes and improvements are possible, which fall within the scope of the claimed disclosure.

Claims (10)

1. A comprehensive ship accident space-time hot spot dynamic analysis method is characterized by comprising the following steps:

receiving ship accident data, and performing data filtering and data conversion on the ship accident data to obtain ship accident handling data, wherein the ship accident data comprises ship type, event abstract, total tonnage, class society, accident coordinates, accident occurrence time, traffic flow data of casualties and areas where the ship accidents are located in the ship accidents;

positioning the ship accident based on the ship accident handling data to obtain the geographical position distribution of the ship accident, and performing density analysis, space autocorrelation analysis and three-dimensional space-time hot spot analysis on the ship accident based on the geographical position distribution of the ship accident;

and respectively carrying out parameter adjustment on the density analysis, the space autocorrelation analysis and the three-dimensional space-time hot spot analysis, and obtaining a ship accident space-time hot spot output result through multiple screening.

2. The comprehensive ship accident space-time hot spot dynamic analysis method according to claim 1, wherein the data filtering comprises the steps of filtering out accident data without coordinates in ship accident data, clearing abnormal data and correcting error data, wherein the data is converted into the longitude and latitude coordinate data of the ship accident in a bisection format, and the longitude and latitude coordinate data is converted into decimal form identifiable by geographic information system software.

3. A comprehensive space-time hot spot dynamic analysis method for marine accidents according to claim 1, wherein the density analysis comprises; nuclear density estimation and point density analysis are used to determine the density level and risk boundaries of marine accidents.

4. A comprehensive space-time hot spot dynamic analysis method for ship accidents according to claim 3, wherein the calculation formula of the kernel density estimation is:

wherein f (x, y) is the estimated density at the (x, y) position, s is the number of positions, h is the bandwidth, K is the kernel function, lambda _i Distance from position (x, y) to the i-th position;

the calculation formula of the dot density analysis F is:

wherein ρ is a neighborhood radius, N (ρ) is the number of ship accidents in a circle with the center of the unit k as the center, the neighborhood radius is ρ, and M is the ship traffic density.

5. A comprehensive space-time hot spot dynamic analysis method for marine vessel accidents according to claim 1, wherein the spatial autocorrelation analysis comprises global spatial autocorrelation analysis and local spatial autocorrelation analysis.

6. The comprehensive ship accident space-time hot spot dynamic analysis method according to claim 5, wherein the global space autocorrelation analysis I has a calculation formula:

wherein n is the total number of accidents in the investigation region, x _i And x _j Representing the i-th and j-th spatial observations respectively,w _ij representing elements of a spatial binary adjacency matrix, the weights being neighborhood relations existing between a position i and its neighboring position j, W representing W _ij Sum of all accidents in (a),>

local spatial autocorrelation analysisThe calculation formula is as follows:

wherein S is x _j Is set in the standard deviation of (2),local spatial autocorrelation analysis classifies the degree of accident aggregation into seven categories, hot or cold spots with 90%, 95% and 99% confidence levels, and spots without significant aggregation.

7. The comprehensive space-time hotspot dynamic analysis method for ship accidents according to claim 1, wherein the three-dimensional space-time hotspot analysis comprises: emerging hot spot analysis, spatiotemporal hot spot trend analysis, and spatiotemporal kernel density estimation analysis.

8. The comprehensive ship accident space-time hot spot dynamic analysis method according to claim 7, wherein the emerging hot spot analysis can explain and classify accident hot spots in more detail and can also present changes of different types of hot spots, the emerging hot spot analysis divides the results into hot spots and cold spots which are added, continuous, reinforced, permanent, reduced, scattered, oscillated and historical, the space-time hot spot trend analysis can identify accident hot spot change trend with statistical significance, the space-time hot spot trend analysis divides the trend of z-score of hot spots or cold spots to seven types with rising or falling trend with 90%, 95% and 99% confidence level, and no significant trend;

the calculation formula of the space-time kernel density estimation is as follows:

wherein,for density estimation at location (x, y, t), the x-and y-dimensions are spaceThe dimension t is time, ds is spatial bandwidth, dt is temporal bandwidth, d _s ² d is the density estimate +.>The density value obtained by multiplying n is expressed by the number of events per unit space time, and the calculation formulas of the kernel functions Ls and Lt are as follows:

where u is the difference between longitude xi and xi+1, v is the difference between latitude yi and yi+1, and p is the difference between time ti and ti+1.

9. The comprehensive ship accident space-time hot spot dynamic analysis method according to claim 1, wherein the process of obtaining the ship accident space-time hot spot output result is as follows:

in the parameter adjustment process of the nuclear density estimation, the result of the nuclear density estimation is jointly influenced by a nuclear function K, a bandwidth h and a unit size, the bandwidth influences the smoothness of a density surface, the unit size influences the roughness of a generated surface, and a group of nuclear density estimation output results which are most in line with a ship accident is finally obtained through multiple screening;

in the parameter adjustment process of the point density analysis, the result of the point density analysis is jointly influenced by the neighborhood radius rho and the ship traffic density M, and a group of output results which most accord with the ship accident point density analysis is finally obtained through multiple screening;

in the parameter adjustment process of creating a space-time cube, emerging space-time hot spot analysis, space-time hot spot trend analysis and space-time kernel density estimation, the analysis result is jointly influenced by the time and space of a space-time box, and multiple times of screening are needed, so that a group of analysis output results which are most in line with the space-time hot spot of the ship accident are finally obtained;

and combining the estimated output result of the nuclear density of the ship accident, the analysis output result of the most-accordant ship accident point density and the analysis output result of the most-accordant ship accident space-time hot spot to obtain the ship accident space-time hot spot output result.

10. A comprehensive space-time hot spot dynamic analysis system for ship accidents, comprising:

and a data processing module: the system comprises a ship accident data receiving module, a ship accident data processing module and a ship accident data processing module, wherein the ship accident data receiving module is used for receiving the ship accident data, carrying out data filtering and data conversion on the ship accident data to obtain ship accident processing data, and the ship accident data comprise ship type, event abstract, total tonnage, class society, accident coordinates, accident occurrence time, casualties and traffic flow data of a region where the ship accident is located;

and an accident analysis module: the system is used for positioning the ship accident based on the ship accident handling data to obtain the geographical position distribution of the ship accident, and performing density analysis, spatial autocorrelation analysis and three-dimensional space-time hot spot analysis on the ship accident based on the geographical position distribution of the ship accident;

and a result output module: the method is used for respectively carrying out parameter adjustment on density analysis, space autocorrelation analysis and three-dimensional space-time hot spot analysis, and obtaining a ship accident space-time hot spot output result through multiple screening.