The invention relates to a method and a device for obtaining crowd distribution characteristics

Abstract

The embodiment of the specification provides a method and a device for acquiring crowd distribution characteristics. The method comprises the following steps: inputting a distribution scene image of acrowd scene to be analyzed into a density distribution recognition model obtained in advance for image recognition; The density distribution recognition model is trained according to each sample distribution scene image and corresponding density distribution truth value map, The image of each sample distribution scene includes images captured by different camera shooting parameters, and the density distribution truth value map is obtained according to the scene depth of each human body in the sample distribution scene image. Obtaining a crowd density distribution map output from the density distribution identification model, the crowd density distribution map being used to represent a crowd density distribution status of the crowd scene. The method of this instruction has strong anti-interference ability, and the population distribution characteristics obtained are more accurate.

Description

Technical field [0001] The present disclosure relates to the field of computer technology, in particular to a method and device for obtaining population distribution characteristics. Background technique [0002] With the rapid population growth, there are more and more security incidents caused by crowd gatherings, which has gradually increased government departments’ monitoring of crowds, for example, in some outdoor public scenes (such as squares, scenic spots, downtown blocks, etc.) ), it is more necessary to strengthen the monitoring and evaluation of population distribution. Among them, crowd distribution characteristics such as crowd density distribution and crowd number statistics can be used as an important reference for assessing the risk of crowd gathering. The camera can be used to capture the distribution scene image of the crowd distribution in a public scene, and analyze the distribution scene image to obtain the public scene The distribution characteristics of th...

AI Technical Summary

Problems solved by technology

However, this analysis method is prone to false detection due to the interference of scene factors. For example, the color of human hair is sensitive to the influence of light changes, which may lead to the failure to extract the head area, poor anti-interference ability, and needs to be combined with the angle and height of the camera, etc. parameters to analyze, the processing efficiency is low

Images