The purpose of the research is to develop the automatic analysis of the crowd density, the main movement direction, its nationality, gender, age, etc.

The general approach

Video processing and features selection

Automatic analysis of crowd characteristics on video

Solve scientific problems

  • Evaluation of the crowd characteristics.
  • Analysis of existing methods for the crowd detection on video and calculate its characteristics.
  • Development of methods and techniques for the crowd detection and analysis of its characteristics.
  • Comparative analysis of the developed method.

Scientific novelty of the current research is in the unique analysis of methods for finding and calculating crowd characteristics, crowd detection algorithms and an integrated approach to crowd analysis.

The practical value

In real-time surveillance system it is difficult for an operator to assess crowd characteristics. Existing systems can only evaluate one or two features.

The developing technology will be able to undertake a complex analysis of crowd behavior based on its characteristics and does not require an operator.

Data that is provided by automatic analysis of crowd characteristics at any moment can be used to search for behavioral models of the crowd. While identifying specific behavioral models the system will alert the operator.

Automatic analysis includes an assessment of crowd density, the main direction of movement, composition by sex, age and nationality.
Technology can be used in the following areas:

  • surveillance and control;
  • experimental sociological studies;
  • ensuring security during mass events (festivals, concerts, meetings, etc.);
  • demand analysis, etc.

Фотография толпы

tolpa 4