A detection method for bad driving behavior frequent areas based on satellite positioning data

Abstract

The invention discloses a method for detecting areas where bad driving behavior frequently occurs based on satellite positioning data. The method comprises the following steps: (1) projecting the satellite positioning data into a plane coordinate system; (2) calculating the local extension direction of each sample; (3) Calculate the rectangular neighborhood and density of each sample point; (4) Expand the cluster with the core point as the starting point; (5) Calculate the boundary according to the obtained cluster to obtain the frequent occurrence area of ​​bad driving behavior. The present invention can effectively adapt to data sets of different scales without manually setting parameters, and considers that the distribution of bad driving behaviors is mostly band-shaped, and the detection result is better than the traditional density clustering method.

Description

technical field [0001] The invention relates to a density clustering method, in particular to a spatial density clustering method based on vehicle satellite positioning data. Background technique [0002] Safety issues during vehicle driving have always been the focus of social concern, and bad driving behavior is usually the main cause of traffic accidents. At this stage, based on various data generated during the driving of the vehicle, there are already a large number of methods to detect the occurrence of bad driving behavior of the vehicle. Aiming at the location information of the detected bad driving behaviors, it is of great significance to the traffic safety management to study the distribution law in space (mainly to study which road sections have high frequency of bad driving behaviors). [0003] Bad driving behavior can be represented by satellite positioning data. Each pair of latitude and longitude coordinates represents one bad driving behavior at the curren...

AI Technical Summary

Problems solved by technology

First of all, DBSCAN needs to manually set two parameters, namely the neighborhood radius and density threshold, which need to be adjusted according to the scale of the data set; secondly, the effect of DBSCAN clustering is related to the distance measurement method between data points , the existing distance measurement methods are not suitable for the band-shaped characteristics of the satellite positioning data generated by the vehicle; finally, DBSCAN cannot determine the boundary of the frequent occurrence area

Images