Method and system for identifying black spot area of road traffic accident

Abstract

The invention discloses a road traffic accident black spot area identification method and system, and relates to the technical field of data processing. The method comprises the following steps: collecting total historical road traffic accident data; performing first-order division on the target traffic area based on a nine-grid correlation positioning method, and outputting a first-order space grid; performing coding mapping on the total historical road traffic accident data according to the first-order space grid to generate a road traffic accident coding mapping library; an accident black spot judging device is constructed, the accident black spot judging device calls the road traffic accident code mapping library to carry out accident number recognition, and a black spot grid is output; and carrying out nine-grid grid correlation merging analysis by taking the black point grid as a central grid, outputting a merged black point grid, and carrying out marking visual display on the merged black point grid. The technical problems of inaccurate identification and low efficiency of the traffic accident black spot area in the prior art are solved, and the technical effect of improving the identification precision and efficiency of the traffic accident black spot area is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly relates to a method and system for identifying black spots in road traffic accident areas. Background Art

[0002] In the prior art, the following problems generally exist in accident black spot identification methods: on the one hand, traditional manual analysis methods are inefficient and are easily affected by subjective factors, resulting in inaccurate results; on the other hand, existing methods lack sufficient consideration of spatio-temporal aggregation factors and cannot accurately reflect the accident distribution law in complex road networks. In addition, the black spot identification accuracy of traditional methods is relatively low, usually relying on rough data aggregation and unable to be accurate to specific road sections or intersections. At the same time, existing methods are mostly based on single-dimensional data and lack comprehensive analysis of multi-dimensional factors, resulting in the inability to deeply explore the root causes of accidents and regional characteristics. Finally, the data update frequency of traditional methods is relatively low, making it difficult to identify new accident black spots in a timely manner and unable to effectively address traffic safety hazards. Summary of the Invention

[0003] This application provides a method and system for identifying black spots in road traffic accident areas, which solves the technical problems of inaccurate identification and low efficiency in identifying black spots in road traffic accident areas in the prior art.

[0004] In view of the above problems, this application provides a method and system for identifying black spots in road traffic accident areas.

[0005] In the first aspect of this application, a method for identifying black spots in road traffic accident areas is provided. The method includes: collecting all historical road traffic accident data; performing a first-order division on a target traffic area based on the nine-square correlation positioning method to output a first-order spatial grid, where the first-order spatial grid includes first-order setting parameters , and are the length and width of the first-order spatial grid; performing encoding mapping on the all historical road traffic accident data according to the first-order spatial grid to generate a road traffic accident encoding mapping library; constructing an accident black spot determiner, where the accident black spot determiner calls the road traffic accident encoding mapping library to identify the number of accidents and outputs a black spot grid; performing nine-square grid correlation merging analysis with the black spot grid as the central grid to output a merged black spot grid, and performing marked visualization display on the merged black spot grid.

[0006] In the second aspect of this application, a system for identifying black spots in road traffic accident areas is provided. The system includes: a data collection module: collecting all historical road traffic accident data; a region division module: performing a first-order division on a target traffic area based on the nine-square correlation positioning method to output a first-order spatial grid, where the first-order spatial grid includes first-order setting parameters , and are the length and width of the first-order spatial grid; Encoding mapping module: encoding and mapping the full amount of historical road traffic accident data according to the first-order spatial grid to generate a road traffic accident encoding mapping library; Accident black spot determination module: constructing an accident black spot determiner, and the accident black spot determiner calls the road traffic accident encoding mapping library to identify the number of accidents and output a black spot grid; Relevance analysis module: performing nine-grid relevance merging analysis with the black spot grid as the central grid, outputting a merged black spot grid, and performing marked visualization display on the merged black spot grid.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: First, collect the full amount of historical road traffic accident data. Then, based on the nine-grid correlation positioning method, perform first-order division on the target traffic area, and output a first-order spatial grid. The first-order spatial grid includes first-order setting parameters , and are the length and width of the first-order spatial grid. Further, encode and map the full amount of historical road traffic accident data according to the first-order spatial grid to generate a road traffic accident encoding mapping library. Then, construct an accident black spot determiner, and the accident black spot determiner calls the road traffic accident encoding mapping library to identify the number of accidents and output a black spot grid. Finally, perform nine-grid relevance merging analysis with the black spot grid as the central grid, output a merged black spot grid, and perform marked visualization display on the merged black spot grid. It solves the technical problems of inaccurate identification and low efficiency of accident black spot areas in the prior art, and achieves the technical effects of improving the identification accuracy and efficiency of accident black spot areas. Description of the Drawings

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0009] Figure 1 It is a schematic flowchart of a method for identifying accident black spot areas provided by an embodiment of this application; Figure 2 It is a schematic structural diagram of a system for identifying accident black spot areas provided by an embodiment of this application.

[0010] Description of the reference numerals: Data acquisition module 11, area division module 12, encoding mapping module 13, accident black spot determination module 14, relevance analysis module 15. Specific Embodiment

[0011] By providing a method and system for identifying black spot areas of road traffic accidents, the present application solves the technical problems of inaccurate identification and low efficiency in identifying black spot areas of traffic accidents in the prior art.

[0012] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0013] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server including a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products, or devices.

[0014] Embodiment 1, as Figure 1 shown, the present application provides a method for identifying black spot areas of road traffic accidents, wherein the method includes:

[0015] Collect all historical road traffic accident data. The all historical traffic accident data covers all dimensions from the accident occurrence to the processing result, specifically including the alarm source (such as telephone alarm, automatic alarm of monitoring equipment), processing method (online processing, offline processing), acceptance department, management department, weather condition (such as sunny, rainy, foggy, snowy), location type (such as highway, urban arterial road, rural road), accident classification (such as minor accident, general accident, major accident, extremely serious accident), accident form (such as rear-end collision, side collision, rollover, rolling, collision with pedestrians, collision with obstacles, etc.), processing result, accident type (vehicle accident, vehicle-person accident), accident violation (such as speeding, running a red light, drunk driving, fatigue driving, failure to give way to pedestrians, illegal lane change, etc.), longitude and latitude positioning information (longitude and latitude coordinates of the accident occurrence location), vehicle information (such as car, large truck, bus, motorcycle, etc.), personnel information (driver information, passenger information), etc. By collecting all historical road traffic accident data, a complete data foundation is established to provide support for subsequent spatial grid division, black spot identification and analysis of accidents.

[0016] Based on the nine-square correlation positioning method, perform a first-order division on the target traffic area and output first-order spatial grids, where the first-order spatial grids include first-order set parameters , and are the length and width of the first-order spatial grid.

[0017] The nine-grid correlation positioning method is a positioning method that divides a spatial area into nine small grids, similar to the layout of a nine-grid. Each grid represents a smaller geographical area. Through the nine-grid correlation positioning method, precise spatial analysis can be carried out within a larger area to find the aggregation area where accidents occur.

[0018] Based on the nine-grid correlation positioning method, the target traffic area (such as a certain city or section) is initially divided in the way of a nine-grid, and multiple first-order spatial grids are output. Each grid represents a certain geographical area. Further, the first-order spatial grid refers to the first layer of spatial unit after division. The length and width of the first-order spatial grid are respectively set as and .

[0019] Furthermore, after outputting the first-order spatial grid, the method further includes: performing traffic feature analysis on the target traffic area to obtain the characteristic vector values corresponding to each grid area in the first-order spatial grid; judging according to the magnitudes of the characteristic vector values corresponding to each grid area, and extracting N marked grids with characteristic vector values greater than the preset characteristic vector value; performing second-order division on the N marked grids based on the nine-grid correlation positioning method to output second-order spatial grids, and the second-order spatial grids include second-order settings , and are the length and width of the second-order spatial grid, and ; performing coding mapping on the full amount of historical road traffic accident data according to the second-order spatial grid, and updating the road traffic accident coding mapping library.

[0020] After outputting the first-order spatial grid, traffic characteristics are further analyzed to refine the spatial grid division, map accident data to smaller areas, and then more accurately identify black spots of traffic accidents. Specifically, traffic characteristics of each first-order spatial grid in the target traffic area are analyzed, including but not limited to accident occurrence frequency, traffic flow, road conditions, weather conditions, etc.; by analyzing historical traffic accident data, a characteristic vector value is generated for each first-order grid area. The characteristic vector value is usually a quantization result calculated based on a series of traffic data (such as the number of accidents, traffic flow, etc.). This value represents the traffic condition of the area. For example, an area with frequent accidents will have a larger characteristic vector value. According to the size of the characteristic vector value corresponding to each grid area, they are sorted and N marked grids with characteristic vector values greater than a preset threshold are selected. These grid areas are areas with a high accident occurrence frequency and represent key areas for traffic management. After identifying high-risk areas, second-order division of these marked grids is performed based on the nine-square correlation positioning method, that is, the nine-square grid method is used to further subdivide these grids to improve the accuracy of accident-prone areas; different from the first-order division, the size of the second-order spatial grid is smaller, and the length and width are respectively represented by and and , that is, the second-order spatial grid is more refined than the first-order spatial grid. According to the division of the second-order spatial grid, all historical traffic accident data is mapped to the new second-order grid, that is, each accident data will be assigned to its corresponding second-order spatial grid; at the same time, the road traffic accident coding mapping library is updated to store the mapping relationship between all accident data and the second-order spatial grid.

[0021] Encode and map all historical road traffic accident data according to the first-order spatial grid to generate a road traffic accident coding mapping library.

[0022] Map historical traffic accident data to the previously defined first-order spatial grid to generate a road traffic accident coding mapping library. The road traffic accident coding mapping library associates each accident with its corresponding spatial grid position and forms a database for subsequent retrieval and analysis. Specifically, the geographical coordinates (latitude and longitude) of each accident are converted into grid coordinates relative to the first-order spatial grid, that is, it is determined which grid the accident occurred in, and a unique identifier (ID) is assigned to each grid. Combining the specific time of the accident occurrence, a unique code containing the grid ID and timestamp is generated for each accident; traverse all historical road traffic accident data, and according to the coding mapping rules, map each accident to the corresponding first-order spatial grid, record its unique code, and store the mapping result in the road traffic accident coding mapping library.

[0023] Optionally, the longitude and latitude information is converted into grid numbers using the following formula to achieve a grid-based representation of the spatial region; Longitude number = Earth's equatorial circumference * COS(angle between the current latitude and the equator) * longitude value / (360 * grid length); Latitude number = Earth's meridian circumference * current latitude / (360 * grid width); Grid number = Longitude number - Latitude number.

[0024] Taking 50 meters × 50 meters as an example, the numbers of a certain grid and its 8 surrounding grids are as follows in the table:

[0025] Construct an accident black spot determiner, which calls the road traffic accident coding mapping library to identify the number of accidents and outputs the black spot grids.

[0026] The accident black spot determiner can call the generated road traffic accident coding mapping library above, identify black spots according to the number of traffic accidents occurring, and output the black spot grids, where the black spot refers to the area with a relatively high frequency of traffic accidents.

[0027] Furthermore, the accident black spot determiner calls the road traffic accident coding mapping library to identify the number of accidents and outputs the black spot grids. The method includes: The accident black spot determiner includes a first determination channel and a second determination channel. The first determination channel is an absolute value method determination channel, and the second determination channel is a relative value method determination channel. The first determination channel and the second determination channel are connected in sequence; Based on the first determination channel, the number of accidents in the road traffic accident coding mapping library is calculated to obtain the initial black spot grids greater than the first preset accident number threshold; Based on the second determination channel, the black spot grids greater than the second preset accident number threshold are output from the initial black spot grids.

[0028] The accident black spot detector consists of a first determination channel and a second determination channel. Among them, the first determination channel uses the absolute value method for determination, and the second determination channel uses the relative value method for determination. The two determination channels are connected in sequence to process and screen traffic accident data in turn. Based on the first determination channel (absolute value method), the accident black spot detector calculates the number of accidents for the data in the road traffic accident coding mapping library; by counting the total number of traffic accidents in each grid area, if the number of accidents in a certain grid exceeds the set first preset accident number threshold, then this grid is identified as an initial black spot grid. The absolute value method is based on the absolute number of accidents, that is, directly counting the number of accident occurrences to identify areas with a higher accident occurrence frequency. Based on the second determination channel (relative value method), grids that meet the second preset accident number threshold are further screened from the already identified initial black spot grids; different from the first determination channel, the second determination channel uses the relative value method for judgment. By comparing the relative density of each grid with the set second preset accident number threshold, grids with a higher accident density are screened out, thereby excluding some accidental grids that do not conform to the overall traffic accident density, ensuring that the output black spot grids have higher accuracy and reliability. Through the sequential processing of these two determination channels, the accident black spot detector can first screen out potential black spot areas according to the number of accidents, and then further optimize and accurately identify them through the relative evaluation of accident density. This dual screening mechanism can effectively identify areas with a high traffic accident occurrence frequency and a large accident density, and thus output the final black spot grids.

[0029] Furthermore, the second preset accident number threshold is the average network accident ratio of the dynamic grid area; the dynamic grid area is obtained by expanding the grid based on the first step length with the currently calculated initial black spot grid as the center, and the first step length is the number of expanded grids.

[0030] The second preset accident number threshold is not a fixed value, but is dynamically calculated according to the average network accident ratio of the dynamic grid area. Specifically, calculate the total number of accidents in each dynamic grid area, and then compare this total number with the number of grids in the area to obtain the average accident ratio of this area. This ratio is used as the standard for judging whether the current black spot area is serious. If the accident density (i.e., the relative ratio) of a certain grid area is higher than this average accident ratio, then this grid will be marked as the final black spot grid.

[0031] After identifying the initial black dot grid, in order to more comprehensively evaluate the accident density in the black dot area, it is necessary to also consider other grids around the initial black dot grid. This area is called the dynamic grid area, and the size of the dynamic grid area is controlled by the first step length, which determines the number of expanded grids. For example, if the first step length is set to 2, starting from the initial black dot grid, two layers of grids around it are expanded to form a larger grid area as the dynamic grid area.

[0032] The second preset accident quantity threshold is not a fixed value, but is dynamically calculated according to the average network accident ratio of the dynamic grid area. Specifically, calculate the total number of accidents in each dynamic grid area, and then compare this total with the number of grids in the area to obtain the average accident ratio of the area. This ratio is used as the criterion for judging whether the current black dot area is serious. If the accident density (i.e., the relative ratio) of a certain grid area is higher than this average accident ratio, then this grid will be marked as the final black dot grid.

[0033] Furthermore, the first preset accident quantity threshold and the second preset accident quantity threshold are dynamically updated through a time series dynamic update network layer. The method includes: Training the time series dynamic update network layer, which is obtained by training the ARIMA model until convergence. The training data of the time series dynamic update network layer includes weather labels, holiday labels, traffic change indicators, and historical random deviations; predicting the accident trend of the full amount of historical road traffic accident data according to the time series dynamic update network layer, and outputting the predicted accident growth rate; dynamically updating the first preset accident quantity threshold and the second preset accident quantity threshold according to the predicted accident growth rate.

[0034] The first preset accident quantity threshold and the second preset accident quantity threshold are dynamically updated through a time series dynamic update network layer. A time series model can be trained and time series data can be used to predict the trend of traffic accidents, thereby dynamically adjusting the accident quantity threshold. Specifically, the time series dynamic update network layer is trained through an ARIMA model to obtain the prediction ability of accident trends; the ARIMA (AutoRegressive Integrated Moving Average) model is a commonly used time series analysis model suitable for trend prediction of time series data; by training this model, the time series changes of traffic accident data can be modeled and the accident trends in a future period can be predicted. During the training process, the training data used by the time series dynamic update network layer includes weather labels (the impact of different weather conditions on traffic accidents, such as rainy days, snowy days, haze weather, etc.), holiday labels (special patterns of traffic flow and accident occurrence during holidays or specific vacation periods), traffic flow change indicators (the change trend of traffic flow, such as the flow changes during morning rush hours and evening rush hours), and historical random deviations (the random volatility in historical data, considering the volatility differences in different years, months, or days). After the model converges and the training is completed, the time series dynamic update network layer can predict the accident trend for the full volume of historical road traffic accident data according to the training results and output the predicted accident growth rate. The accident growth rate represents the trend of accident occurrence in a future period, that is, the change speed of the predicted traffic accident quantity.

[0035] According to the predicted accident growth rate, the system dynamically updates the first preset accident quantity threshold and the second preset accident quantity threshold according to the following formulas. First preset accident quantity threshold update formula: , where is the initial accident quantity threshold, is the influence coefficient controlling the impact of the accident growth rate on the threshold update, represents the predicted accident growth rate. Second preset accident quantity threshold update formula: , where is the initial second preset accident quantity threshold, is the influence coefficient controlling the impact of the accident growth rate on the second threshold update, represents the accident growth rate.

[0036] Through the training of the ARIMA model, the time series dynamic update network layer can predict the accident growth trend in real time and dynamically update the first preset accident quantity threshold and the second preset accident quantity threshold according to the predicted accident growth rate.

[0037] Taking the black dot grid as the central grid, perform nine-square grid correlation merging analysis, output the merged black dot grid, and mark and visually display the merged black dot grid.

[0038] Taking the identified black dot grid as the center, perform nine-square grid correlation merging analysis. By analyzing the correlation of the grids around the central black dot, further expand the scope of black dot identification, and obtain the merged black dot grid, which can more comprehensively reveal the concentrated areas of traffic accidents.

[0039] Furthermore, taking the black dot grid as the central grid to perform nine-square grid correlation merging analysis, the method includes: According to the accident black dot detector, grids with a grid identifier greater than or equal to 1 / 2 of the first preset accident quantity threshold and less than the first preset accident quantity threshold are marked as buffer grids, and grids with a grid identifier less than 1 / 2 of the first preset accident quantity threshold are marked as low-incidence grids; obtain the first-order nine-square grid with the black dot grid as the central grid. If the eight adjacent grids in the first-order nine-square grid include the black dot grid or the buffer grid, obtain a first-order merging instruction, and perform nine-square grid correlation merging according to the first-order merging instruction to output the merged black dot grid.

[0040] Specifically, using the accident black spot detector, analyze the number of accidents in each grid area and classify the grids according to the accident frequency; for those grids where the number of accidents is greater than or equal to 1 / 2 of the first preset accident number threshold and less than the first preset accident number threshold, they are marked as buffer grids. The accident occurrence frequency in these grids is close to that of the black spot grids but has not reached the high-risk standard; for grids where the number of accidents is less than 1 / 2 of the first preset accident number threshold, they are marked as low-incidence grids. The accident occurrence frequency in these areas is relatively low, with relatively small risks, and they are generally not regarded as black spot areas. Taking the black spot grid as the central grid, conduct a nine-grid correlation merging analysis. By obtaining the first-order nine-grid, that is, including the central black spot grid and its eight adjacent grids around it, the accident risks of these grids can be further analyzed. The relationship between the frequent accident situation of the central grid and its surrounding grids will determine whether merging is required; for the eight adjacent grids in the first-order nine-grid, if at least one of them is a black spot grid or a buffer grid, it is considered that these grids are closely related to the black spot area and have a certain correlation. These areas have relatively high accident risks and meet the conditions for expanding the black spot area. At this time, the system will generate a first-order merging instruction to indicate the merging of these grids; if the adjacent grids contain low-incidence grids, the merging condition is not met and no merging is performed; according to the generated first-order merging instruction, the system performs the nine-grid correlation merging, merges the grids that meet the merging conditions, and outputs the merged black spot grids. These merged grids show a wider area with frequent accidents, and these merged black spot grids are not necessarily regular rectangles. It may also be an irregular shape, which mainly depends on the density of accident occurrences within the grid and its geographical adjacency relationship. The shape and size of the merged area will vary according to the actual accident distribution. In addition, the number of grids merged in each order of merging operation is limited. The number of merges in each order is less than or equal to 8, that is, in each round of the merging process, at most 8 adjacent grids can be merged into the black spot area to ensure the accuracy and rationality of the merged area. By taking the black spot grid as the central grid and using the nine-grid correlation merging analysis, the system can merge adjacent areas according to the accident frequency and generate new merged black spot grids. This method can identify a larger range of high-risk areas, help the traffic management department take targeted safety improvement measures, and improve the accuracy and efficiency of traffic safety management.

[0041] Furthermore, after outputting the merged black spot grids, the method further includes: Obtain the second-order nine-grid with the merged black spot grid as the central grid; if the eight adjacent grids in the second-order nine-grid include the black spot grid or the buffer grid, obtain the second-order merging instruction, perform the nine-grid correlation merging according to the second-order merging instruction, output the second-order merged black spot grids, and perform marked visualization display on the second-order merged black spot grids.

[0042] After outputting the merged black dot grid, further analyze the extended area of the merged black dot grid to identify a wider high-risk area. Specifically, using the merged black dot grid as the central grid, obtain a second-order nine-square grid. The second-order nine-square grid is a larger grid centered on the merged black dot grid and includes the eight adjacent grids around it. Next, judge the eight adjacent grids in the second-order nine-square grid: If one or more grids in the second-order nine-square grid belong to the black dot grid or the buffer grid, it is considered that this area has a relatively high accident risk and meets the conditions for further merging; in this case, the system will generate a second-order merging instruction indicating that these grids can be merged to form a wider black dot area. By performing the nine-square grid correlation merging, the system will merge the eligible grids and output the second-order merged black dot grid. Among them, the merged black dot grid is not necessarily a regular rectangle and may also present an irregular shape, depending on the actual situation of the accident distribution and the grid form during the merging process. This merging process expands the identification range of high-accident areas, thus helping the traffic management department accurately identify potential high-risk areas. Finally, the generated second-order merged black dot grid will be displayed through marked visualization, using different colors or symbols to mark these high-risk areas on the map, facilitating traffic management personnel to quickly identify and take corresponding safety improvement measures.

[0043] In summary, the embodiments of the present application have at least the following technical effects: First, collect all historical road traffic accident data. Then, based on the nine-square correlation positioning method, perform a first-order division on the target traffic area and output the first-order spatial grid. The first-order spatial grid includes the first-order setting parameters , and which are the length and width of the first-order spatial grid. Further, encode and map all historical road traffic accident data according to the first-order spatial grid to generate a road traffic accident coding and mapping library. Then, construct an accident black dot determiner. The accident black dot determiner calls the road traffic accident coding and mapping library to identify the number of accidents and outputs the black dot grid. Finally, perform nine-square grid correlation merging analysis with the black dot grid as the central grid, output the merged black dot grid, and perform marked visualization display on the merged black dot grid. This solves the technical problems of inaccurate identification and low efficiency of accident black dot areas in the prior art, and achieves the technical effects of improving the identification accuracy and efficiency of accident black dot areas.

[0044] Embodiment 2, based on the same inventive concept as the method for identifying a road traffic accident black dot area in the foregoing embodiment, as Figure 2 shown, the present application provides a system for identifying a road traffic accident black dot area, where the system includes: Data acquisition module 11: Collect all historical road traffic accident data; Region division module 12: Perform first-order division on the target traffic region based on the nine-square correlation positioning method, and output first-order spatial grids, where the first-order spatial grids include first-order setting parameters , and are the length and width of the first-order spatial grid; Coding mapping module 13: Perform coding mapping on the all historical road traffic accident data according to the first-order spatial grids to generate a road traffic accident coding mapping library; Accident black spot determination module 14: Construct an accident black spot determiner, which calls the road traffic accident coding mapping library to identify the number of accidents and outputs black spot grids; Correlation analysis module 15: Perform nine-square grid correlation merging analysis with the black spot grids as the central grids, output merged black spot grids, and perform marked visualization display on the merged black spot grids.

[0045] Furthermore, the region division module 12 is used to execute the following method: Perform traffic feature analysis on the target traffic region to obtain the characteristic vector values corresponding to each grid region in the first-order spatial grids; Make judgments according to the magnitudes of the characteristic vector values corresponding to each grid region, and extract N marked grids with characteristic vector values greater than the preset characteristic vector value; Perform second-order division on the N marked grids based on the nine-square correlation positioning method, and output second-order spatial grids, where the second-order spatial grids include second-order setting parameters , and are the length and width of the second-order spatial grid, and ; Perform coding mapping on the all historical road traffic accident data according to the second-order spatial grids to update the road traffic accident coding mapping library.

[0046] Furthermore, the accident black spot determination module 14 is used to execute the following method: The accident black spot determiner includes a first determination channel and a second determination channel. The first determination channel is an absolute value method determination channel, and the second determination channel is a relative value method determination channel. The first determination channel and the second determination channel are connected in sequence; Calculate the number of accidents in the road traffic accident coding mapping library based on the first determination channel to obtain initial black spot grids with the number of accidents greater than the first preset accident number threshold; Output black spot grids with the number of accidents greater than the second preset accident number threshold from the initial black spot grids based on the second determination channel.

[0047] Furthermore, the accident black spot determination module 14 is used to execute the following method: The second preset accident quantity threshold is the average network accident ratio of the dynamic grid area; the dynamic grid area is centered on the initially calculated initial black spot grid and obtained by expanding the grid based on the first step length, and the first step length is the number of expanded grids.

[0048] Further, the accident black spot determination module 14 is used to execute the following method: Train the time series dynamic update network layer, which is obtained by training the ARIMA model until convergence. The training data of the time series dynamic update network layer includes weather labels, holiday labels, traffic change indicators, and historical random deviations; predict the accident trend of the full amount of historical road traffic accident data according to the time series dynamic update network layer, and output the predicted accident growth rate; dynamically update the first preset accident quantity threshold and the second preset accident quantity threshold according to the predicted accident growth rate.

[0049] Further, the relevance analysis module 15 is used to execute the following method: According to the accident black spot determiner, grids greater than or equal to 1 / 2 of the first preset accident quantity threshold and less than the first preset accident quantity threshold are marked as buffer grids, and grids less than 1 / 2 of the first preset accident quantity threshold are marked as low-incidence grids; obtain the first-order nine-grid centered on the black spot grid. If the adjacent eight grids in the first-order nine-grid include the black spot grid or the buffer grid, obtain the first-order merge instruction, and perform nine-grid relevance merge according to the first-order merge instruction to output the merged black spot grid.

[0050] Further, the relevance analysis module 15 is used to execute the following method: Obtain the second-order nine-grid centered on the merged black spot grid; if the adjacent eight grids in the second-order nine-grid include the black spot grid or the buffer grid, obtain the second-order merge instruction, and perform nine-grid relevance merge according to the second-order merge instruction to output the second-order merged black spot grid, and perform marked visualization display on the second-order merged black spot grid.

[0051] It should be noted that the above sequence of embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above description of specific embodiments of this specification has been made. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0052] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0053] This specification and the accompanying drawings are merely exemplary descriptions of the present application and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A method for identifying black spot areas of road traffic accidents, characterized in that, The method includes: Collecting all historical road traffic accident data; Perform a first-order division on the target traffic area based on the nine-grid correlation positioning method, and output the first-order spatial grid, where the first-order spatial grid includes first-order setting parameters , and are the length and width of the first-order spatial grid; Performing coding mapping on the all historical road traffic accident data according to the first-order spatial grid to generate a road traffic accident coding mapping library; Constructing an accident black spot determiner, which calls the road traffic accident coding mapping library to identify the number of accidents and outputs black spot grids; Taking the black spot grids as the central grids for nine-grid correlation merging analysis, outputting merged black spot grids, and performing marked visualization display on the merged black spot grids.

2. The identification method of a black-spot area of a road traffic accident according to claim 1, wherein After outputting the first-order spatial grid, the method further includes: Performing traffic feature analysis on the target traffic area to obtain the feature vector values corresponding to each grid area in the first-order spatial grid; Judging according to the size of the feature vector values corresponding to each grid area, and extracting N marked grids with feature vector values greater than a preset feature vector value; Perform a second-order division on the N identification grids based on the nine-square correlation positioning method, and output second-order spatial grids, where the second-order spatial grids include second-order setting parameters , and are the length and width of the second-order spatial grid, and ; Performing coding mapping on the all historical road traffic accident data according to the second-order spatial grid to update the road traffic accident coding mapping library.

3. The identification method of a black spot area of a road traffic accident according to claim 1, characterized in that, The accident black spot determiner calls the road traffic accident coding mapping library to identify the number of accidents and outputs black spot grids. The method includes: The accident black spot determiner includes a first determination channel and a second determination channel. The first determination channel is an absolute value method determination channel, and the second determination channel is a relative value method determination channel. The first determination channel and the second determination channel are connected in sequence; Calculating the number of accidents in the road traffic accident coding mapping library based on the first determination channel to obtain initial black spot grids with the number of accidents greater than a first preset accident number threshold; Outputting black spot grids with the number of accidents greater than a second preset accident number threshold from the initial black spot grids based on the second determination channel.

4. The identification method of a black spot area of a road traffic accident according to claim 3, characterized in that The second preset accident number threshold is the average network accident ratio of the dynamic grid area; The dynamic grid area is obtained by expanding the grid based on the first step length with the currently calculated initial black spot grid as the center. The first step length is the number of expanded grids.

5. The method for identifying a black spot area of a road traffic accident according to claim 3, wherein, The first preset accident number threshold and the second preset accident number threshold are dynamically updated through a time series dynamic update network layer. The method includes: Training the time series dynamic update network layer, which is obtained by training the ARIMA model until convergence. The training data of the time series dynamic update network layer includes weather labels, holiday labels, traffic flow change indicators, and historical random deviations; Performing accident trend prediction on the all historical road traffic accident data according to the time series dynamic update network layer, and outputting a predicted accident growth rate; Dynamically updating the first preset accident number threshold and the second preset accident number threshold according to the predicted accident growth rate.

6. The identification method of a black spot area of a road traffic accident according to claim 1, characterized in that, Taking the black spot grids as the central grids for nine-grid correlation merging analysis. The method includes: According to the accident black spot determiner, grids with the number of accidents greater than or equal to 1 / 2 of the first preset accident number threshold and less than the first preset accident number threshold are marked as buffer grids, and grids with the number of accidents less than 1 / 2 of the first preset accident number threshold are marked as low-incidence grids; Obtain the first-order nine-square grid with the black dot grid as the central grid. If the black dot grid or the buffer grid is included in eight adjacent grids of the first-order nine-square grid, obtain the first-order merging instruction, perform the relevance merging of the nine-square grids according to the first-order merging instruction, and output the merged black dot grid.

7. The identification method of a road traffic accident black spot area according to claim 6, characterized in that, After outputting the merged black dot grid, the method further includes: Obtain the second-order nine-square grid with the merged black dot grid as the central grid; If the black dot grid or the buffer grid is included in eight adjacent grids of the second-order nine-square grid, obtain the second-order merging instruction, perform the relevance merging of the nine-square grids according to the second-order merging instruction, output the second-order merged black dot grid, and perform marked visualization display on the second-order merged black dot grid.

8. An identification system for black spots in road traffic accident areas, characterized in that, For implementing a method for identifying black spot areas of road traffic accidents according to any one of claims 1-7, the system includes: Data acquisition module: Collect all historical road traffic accident data; Region division module: Based on the nine-square correlation positioning method, perform a first-order division on the target traffic area and output first-order spatial grids, where the first-order spatial grids include first-order setting parameters , and are the length and width of the first-order spatial grid; Coding mapping module: Perform coding mapping on the all historical road traffic accident data according to the first-order spatial grid to generate a road traffic accident coding mapping library; Accident black spot determination module: Construct an accident black spot determiner, and the accident black spot determiner calls the road traffic accident coding mapping library to identify the number of accidents and output the black dot grid; Relevance analysis module: Perform nine-square grid relevance merging analysis with the black dot grid as the central grid, output the merged black dot grid, and perform marked visualization display on the merged black dot grid.

Images