Fighting behavior detection method based on geographic video

Abstract

The invention discloses a fight behavior detection method based on a geographic video. The method comprises the following steps: firstly detecting a fight behavior in a video image through employing a behavior detection algorithm based on deep learning, and outputting the position information of a behavior main body in an image coordinate system; mapping position information in the image coordinate system to a world coordinate system, and mapping dynamic information obtained by the deep learning model to a static geographic space; and finally, automatically marking the position information of the behavior subject in the remote sensing image to complete visualization. Through fusion and visual design of fight behavior detection and geographic videos, position information in a monitoring picture can be accurately mapped to a geographic space, global mastering of the safety condition of a monitoring scene is facilitated, automatic detection of fight behaviors is achieved, detection information is mapped to the geographic space, and automatic plotting is achieved in a remote sensing image map.

Description

technical field [0001] The invention belongs to the field of computer information technology, relates to the field of deep learning and visual SLAM, in particular to a method for detecting fighting behavior based on geographical video. Background technique [0002] The development direction of the intelligent monitoring system has shifted from simple face recognition and vehicle recognition to more complex binary classification and multi-classification problems. [0003] The interdisciplinary integration of deep learning and surveying and mapping can map the fragmented image space information into geographic space, and construct a hypermedia map that can dynamically perceive the monitoring area, which is helpful for the overall grasp of the area. Using deep learning to process surveillance video can effectively solve the shortcomings of traditional video surveillance systems, automatically identify sudden events in the surveillance area, and equip surveillance cameras with i...

AI Technical Summary

Problems solved by technology

Specifically, today's behavior detection faces two key challenges: 1. It is impossible to detect the fighting behavior in the monitoring scene, and it lacks early warning; 2. With the widespread deployment of surveillance cameras, there is a gap between the picture and the real scene. Complicated correspondence

Images