Real-time detection method for large-scale group aggregation event

Abstract

The invention discloses a real-time detection method for a large-scale group aggregation event. The method comprises the following steps: 1, acquiring time sequence data; 2, calculating a crowd density expected value; 3, calculating a crowd density threshold value; 4, performing crowd density prediction based on the expansion convolutional neural network; and 5, predicting an aggregation event. According to the invention, the method can achieve the effective detection of the number change of crowds at a target position, and carries out the accurate recognition and prediction of the possible gathering events, thereby better carrying out the analysis of a large-scale group activity track, carrying out the early warning of the group events, and maintaining the public safety.

Description

technical field [0001] The invention relates to a method for detecting aggregation events, in particular to a method for detecting large-scale group aggregation events based on crowd density time series data and an expanded convolutional neural network method. Background technique [0002] In recent years, with the continuous development of smart mobile devices and positioning technology, communication base stations have also achieved comprehensive coverage, and communication service providers can more accurately collect and detect user data. The analysis of the number of users in a specific area has great theoretical and practical significance, and is of great value for the study of crowd mobility and urban dynamics. The expansion of cities and the increase in the number of urban populations pose unprecedented challenges to urban management. In terms of urban planning, the identification of large-scale crowd gathering points can optimize urban layout, identify urban hot spo...

AI Technical Summary

Problems solved by technology

However, there are many deficiencies in the aggregation behavior prediction based on social network data. On the one hand, due to the subjectivity of users’ location sharing behavior, the location information provided by social networking sites is incomplete, and the limited number of users on social networks is exacerbated. data incompleteness

On the other hand, the method based on social network data has poor real-time performance, and it is difficult to capture the law of crowd flow in real time.

[0004] There are also some methods to analyze the phenomenon of urban agglomeration through the urban surveillance camera system, and use image processing-based methods for crowd density analysis (An automatic scale-adaptive approach with attention mechanism-based crowd spatial information for crowd counting), but these methods also There are limitations, such as the camera equipment is easily blocked by other obstacles, and the distribution of camera equipment is discrete, it is difficult to capture the overall situation, and it is difficult for the camera system to record events that occur in the wild or at night

[0005] The comprehensive coverage of communication base stations provides a new solution to solve the real-time and integrity problems of user data. At present, there are also some methods based on the data of communication base stations for research on crowd gathering events (Research on the impact of crowd flow on crowd risk in large gathering spots ), but these methods mainly analyze the characteristics of crowd movement, and are less effective in real-time prediction of gathering events

[0006] To sum up, the detection of large-scale group gathering events is of great significance for urban planning and urban security. In the existing research on this problem, a large part of the methods face the problem of incomplete data, and the existing algorithms are very poor. Little consideration is given to the real-time nature of event detection

Images