Abnormal fall behavior detection method based on spatio-temporal motion characteristics

Abstract

The invention relates to a method for detecting abnormal fall behavior based on time-space motion characteristics. Based on image information, the five-point inverted pendulum model that reflects the posture of the human body is refined, and then the standard space-time model of motion that reflects the dual characteristics of time and space is constructed; then based on the standard space-time diagram, the essential characteristics of motion are studied by applying the principle of dynamics. Construct a fixed-axis motion model, realize feature quantification with rotational energy, construct a feature-encoded data set of falling behavior, normal walking or standing state, and conduct training to form a second-class classifier, which is used to classify the processed real-time collected video data Identify and judge whether it is an abnormal falling behavior. The invention adopts a single-view scene, and the algorithm has stronger real-time performance, which avoids the synchronization of multi-view scenes in scene fusion and other problems; it uses rotation energy as a feature, avoids the limitation of morphological features on forward falling posture, and improves The detection is effective, and it has good universality for typical fall postures.

Description

technical field [0001] The invention relates to a video-based abnormal behavior recognition technology, in particular to an automatic detection method for an abnormal fall event in a public place. Background technique [0002] Video surveillance system (CCTV) is a comprehensive application of multimedia technology, computer network, industrial control and artificial intelligence, and is constantly developing in the direction of digitization, system networking and intelligent management. At present, video surveillance is widely used in information acquisition, command and dispatch, security prevention, etc., and can provide various services such as production process control, medical monitoring, distance education, and security of large public facilities. However, there are obvious deficiencies in the current application of closed-circuit television monitoring systems. First of all, due to the limited space in the monitoring room or monitoring center, it is only possible to ...

AI Technical Summary

Problems solved by technology

However, this method is only effective for some fall posture recognition (it is difficult to detect if the shooting angle is parallel or close to the fall direction), and it may fail due to the occlusion of people or objects

Under complex actual scene conditions, the existing abnormal behavior detection method that uses image data to analyze the temporal and spatial characteristics of moving objects can identify and detect abnormal falling postures (which are more confusing with normal walking) in the same direction as the shooting angle. has limitations

Images