CN108765954B - Road traffic safety condition monitoring method based on SNN density ST-OPTIC improved clustering algorithm - Google Patents

Abstract

The invention discloses a road traffic safety condition monitoring method based on SNN density ST-OPTICS improved clustering algorithm, comprising the following steps: S1, collecting vehicle and mobile phone positioning data on a selected road; S2, preprocessing the data, Incorporated into the database; S3, map matching through key positioning information; S4, calculate the SNN density similarity map, and estimate the number of vehicles driving on the selected road according to the similarity map; S5, use ST‑OPTICS clustering algorithm The mobile phone positioning data points are clustered and sorted, and the output clusters are sorted; S6, the passenger object metric value is used as the support degree, and the corresponding vehicle information is stored in the database; S7, the data set satisfying the minimum support degree is obtained for positioning analysis. The invention uses smart phone positioning data, base station positioning data and vehicle-mounted GPS technology to establish a detection model, and intelligently detects the traffic safety status of selected roads.

Description

Road traffic safety monitoring method based on SNN density ST-OPTICS improved clustering algorithm

technical field

The invention relates to the fields of intelligent transportation, data mining and big data processing and analysis, in particular to a road traffic safety condition monitoring method based on SNN density ST-OPTICS improved clustering algorithm.

Background technique

SNN (Shared Nearest Neighbor, Shared Nearest Neighbor) density measures the degree to which a point is surrounded by similar points (about nearest neighbors), and the cluster points found by SNN density-based clustering are all strongly correlated with each other. In the clustering algorithm based on DBSCAN (Density-Based Spatial Clustering of Applications with Noise), the initial parameter values ε Neighborhood and Minpts need to be set manually, and slight differences in parameter values may lead to differences Great clustering results. In order to overcome the disadvantage of using a set of global parameters in clustering analysis, an OPTICS (Ordering Points to identify the clustering structure) clustering analysis method is proposed, which does not explicitly generate a dataset. Instead, generate an augmented cluster ordering based on the parameters ε neighborhood, Minpts, which is a linear table of all analyzed objects and represents the density-based clustering structure of the data.

SUMMARY OF THE INVENTION

In order to solve the defect in the prior art that the vehicle on the selected road needs to be monitored and enforced manually, the present invention proposes a road traffic safety monitoring method based on the SNN density ST-OPTICS improved clustering algorithm, which uses a mobile phone Positioning, base station positioning data and in-vehicle GPS technology can intelligently monitor vehicle traffic safety conditions on selected roads.

The technical scheme adopted in the present invention is:

A road traffic safety condition monitoring method based on SNN density ST-OPTICS improved clustering algorithm, comprising:

S1: Collect GPS positioning data of vehicles driving on the selected road, GPS positioning data and base station positioning data of on-board passenger mobile phones;

S2: Preprocess the data to generate an operable data set and store it in the database;

S3: Map matching with key positioning information including latitude and longitude of actionable datasets;

S4: Calculate the SNN density similarity map using the mobile phone positioning trajectory data points in the operability data set, and automatically determine the number of clusters in the operability data set, so as to estimate the number of vehicles driving on the selected road;

S5: Cluster analysis is performed on the vehicle and mobile phone positioning data points through the ST-OPTICS clustering algorithm, and the corresponding cluster ranking is output, so as to obtain the passenger object measurement value corresponding to each cluster;

S6: take the passenger object metric value as the support degree of the positioning analysis model, take the number of people in the vehicle N as the minimum support degree according to the actual regulations of the selected road, detect the driving vehicle, and store the corresponding vehicle information in the corresponding database;

S7: The data set that satisfies the minimum support degree is used as the basis for the selected road vehicle and mobile phone positioning analysis model.

Preferably, the preprocessing performed on the data in S2 includes comparing data formats, eliminating logically incorrect data, and complementing default data.

Preferably, the eliminating logical error data includes deleting the positioning data with the number of visible satellites less than 4.

Preferably, the complementing the default data includes calculating the position of the trajectory points 30 seconds before and after the signal is missing, using the two center positions before and after the signal is missing as endpoints, and evenly filling the corresponding number of trajectory points according to the spacing.

Preferably, in S6, the number of people on board the vehicle is 2.

Compared with the prior art, the beneficial effects of the present invention are:

The invention innovatively combines the SNN density and the ST-OPTICS clustering algorithm to create a new clustering algorithm, and uses the algorithm combined with the smart phone positioning technology and the vehicle GPS technology to establish a positioning analysis model, which can detect the number of vehicle personnel to be detected. It can save the cost of infrastructure construction and provide a new monitoring method for road law enforcement in the traffic part.

Of course, it is not necessary for any product embodying the present invention to achieve all of the above-described advantages simultaneously.

Description of drawings

FIG. 1 is a flowchart of a road traffic safety condition monitoring method based on an improved SNN density ST-OPTICS clustering algorithm according to an embodiment of the present invention.

Detailed ways

In order to make the objectives, technical solutions and advantages of the present invention clearer, each embodiment of the present invention will be described in detail below with reference to the accompanying drawings.

The following is a detailed description of the vehicle traffic safety monitoring of the HOV lane as an example, but does not limit the protection scope of the present invention.

ST-OPTICS is based on the OPTICS analysis method, using the timestamps of spatiotemporal objects to improve the cluster ordering of data objects. Combining the SNN density with the ST-OPTICS algorithm, a new clustering algorithm is created, starting with the SNN similarity graph for clustering, and building a vehicle and smartphone positioning analysis model based on the algorithm, which mainly includes three steps: The first stage is to calculate the SNN similarity map for the operability data set, and estimate the number of vehicles driving on the HOV lane, that is, the number of clusters that represent the operability data allocation. The second stage uses the ST-OPTICS clustering algorithm. Smartphone positioning data and base station positioning data are clustered, and an augmented cluster ranking is output, representing a linear table of analysis objects. The third stage is to calculate the passenger object metric value corresponding to each cluster according to the ST-OPTICS clustering algorithm based on SNN density. The passenger object measurement is used as the support degree of the positioning analysis model, and the minimum support degree is based on the actual number of vehicles in the HOV lane.

The specific description of the ST-OPTICS clustering algorithm based on SNN density is shown in Table 1:

Table 1 ST-OPTICS clustering algorithm based on SNN density

Referring to Figure 1, a road traffic safety condition monitoring method based on SNN density ST-OPTICS improved clustering algorithm includes the following 7 steps, specifically:

Step 1: Collect GPS positioning data of vehicles driving on the HOV lane and GPS positioning data and base station positioning data of on-board passengers' smartphones. Among them, the collected cell phone base station data is used as auxiliary data for cell phone GPS positioning data to make cell phone positioning more accurate.

Step 2: Preprocess the data to generate an actionable data set and store it in the database. Among them, the preprocessing of the data includes comparing data formats, eliminating logical error data, and complementing default data. Due to the large amount of mobile phone GPS positioning data, uncontrollable changes in the data will occur at various stages of collection, upload and download. The common data format errors are in the form of garbled codes, data repetition, etc.; these data with obvious problems are analyzed by comparison and analysis. delete directly. In addition, because the mobile phone may also have abnormal working states such as jamming and black screen during the data collection process, or due to the occlusion of high-rise buildings, the number of satellites that the mobile phone can receive GPS signals is small, resulting in a large deviation in GPS positioning. Therefore, during data preprocessing, the positioning data with less than 4 visible satellites will be deleted. To supplement the missing data, that is, calculate the position of the trajectory points 30 seconds before and after the signal is missing, take the two center positions before and after the signal is missing as the endpoints, and evenly fill the corresponding number of trajectory points according to the spacing.

Step3: Map matching with key positioning information including the latitude and longitude of the actionable dataset.

Step4: Calculate the SNN density similarity map using the mobile phone positioning trajectory data points in the operability data set, and automatically determine the number of clusters in the operability data set, so as to estimate the number of vehicles driving on the HOV lane. Among them, the SNN (Shared Nearest Neighbor) density similarity map reflects the degree to which a point is surrounded by similar points (about nearest neighbors), and the points in the clusters found by SNN density-based clustering are all strongly correlated with each other. Points in low-density regions of high density generally have relatively high SNN density. The similarity map means that the number of clusters in the operability data set can be automatically determined from the map, and the number of vehicles traveling in the HOV lane can be estimated.

Step5: Perform cluster analysis on the vehicle and mobile phone positioning data points through the ST-OPTICS clustering algorithm, and output the corresponding cluster ranking, so as to obtain the passenger object measurement value corresponding to each cluster.

Step6: Use the passenger object measurement value as the support degree of the positioning analysis model, according to the actual regulations of the HOV lane: the number of vehicles on board is greater than 2 as its minimum support degree, detect the driving vehicle, and store the corresponding vehicle information in the corresponding database.

Step7: Use the data set that satisfies the minimum support degree as the basis for the analysis model of vehicle and mobile phone positioning in the HOV lane.

Compared with the traditional application of infrared thermal imaging technology combined with video monitoring equipment to take pictures, manually identifying the actual number of occupants of vehicles on the road, and relying on video monitoring equipment to automatically capture and take pictures of vehicles entering the HOV lane in violation of regulations, the present invention is innovative. It is proposed to combine SNN density and ST-OPTICS clustering algorithm to create a new clustering algorithm, and use this algorithm to combine smartphone positioning technology and vehicle GPS technology to establish a positioning analysis model, which can detect vehicle personnel on selected roads. , the selected road is not limited to the HOV lane, it is a road selected according to the needs of traffic safety situation detection. The monitoring method based on the positioning analysis model can save the cost of infrastructure construction and provide a new monitoring method for road law enforcement in the traffic part.

The above description is only a preferred embodiment of the present invention, but the protection scope of the present invention is not limited to this. Substitutions should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims (5)

1. a road traffic safety condition monitoring method based on SNN density ST-OPTICS improvement clustering algorithm, is characterized in that, comprises: S1: Collect GPS positioning data of vehicles driving on the selected road, GPS positioning data and base station positioning data of on-board passenger mobile phones; S2: Preprocess the data to generate an operable data set and store it in the database; S3: Map matching with key positioning information including latitude and longitude of actionable datasets; S4: Calculate the SNN density similarity map using the mobile phone positioning trajectory data points in the operability data set, and automatically determine the number of clusters in the operability data set, so as to estimate the number of vehicles driving on the selected road; S5: Cluster analysis is performed on the vehicle and mobile phone positioning data points through the ST-OPTICS clustering algorithm, and the corresponding cluster ranking is output, so as to obtain the passenger object measurement value corresponding to each cluster; S6: take the passenger object metric value as the support degree of the positioning analysis model, take the number of people in the vehicle N as the minimum support degree according to the actual regulations of the selected road, detect the driving vehicle, and store the corresponding vehicle information in the corresponding database; S7: The data set that satisfies the minimum support degree is used as the basis for the selected road vehicle and mobile phone positioning analysis model. 2. the road traffic safety condition monitoring method based on SNN density ST-OPTICS improvement clustering algorithm according to claim 1, is characterized in that, the preprocessing that data is carried out in S2 comprises comparing data format, excluding logical error data and Complete default data. 3 . The road traffic safety condition monitoring method based on SNN density ST-OPTICS improved clustering algorithm according to claim 2 , characterized in that, said eliminating logical error data comprises deleting positioning data whose number of visible satellites is less than 4. 4 . 4. the road traffic safety condition monitoring method based on SNN density ST-OPTICS improved clustering algorithm according to claim 2, it is characterized in that, described complementing default data comprises calculating the track point of each 30 seconds before and after signal missing. Position, take the two center positions before and after the signal is missing as the endpoints, and evenly fill the corresponding number of track points according to the spacing. 5. the road traffic safety condition monitoring method based on SNN density ST-OPTICS improvement clustering algorithm according to claim 1, is characterized in that, described in S6 meets the number of people on board of vehicles N is 2.

Images