Campus security inspection method based on AI vision algorithm

Abstract

The application discloses a campus security inspection method based on an AI vision algorithm, and relates to the field of campus security inspection, and comprises the following steps: deploying a key node vision sensor network covering main roads, public areas and part of the perimeter; extracting personnel identification features in each sampling period to form standardized personnel feature data chains; calculating the risk coefficient of the current position personnel; intelligently broadcasting based on a geographic routing strategy to determine the personnel action trajectory; according to the risk coefficient of the current position personnel, judging the probability of abnormal off-campus of the personnel when the current position reaches the next key node; setting a probability interval and defining a risk label for each interval, and issuing a risk warning according to the risk label. The application has the advantages that the personnel trajectory can be effectively tracked and the probability of abnormal off-campus can be dynamically predicted, so that efficient, accurate and hierarchical warning of abnormal off-campus behavior of campus personnel can be realized.

Images