Method for detecting physical conflict behavior based on low-dimensional spatiotemporal feature extraction and topic modeling

Abstract

The invention belongs to the field of video surveillance technology, and relates to a method for detecting a physical conflict behavior based on low-dimensional spatiotemporal feature extraction and topic modeling. The detection steps are to first define a wordbook, quantize the pixel location of an object, describe the size of a foreground target in a scene, and determine the movement of the foreground pixel. After the above steps, the establishment of the complete wordbook and the corpus establishment are completed. The above calculation method is used to determine a physical conflict behavior. This method combines low-dimensional data feature representation and model-based complex scene analysis, utilizes the change in the location information of a human body in the action to learn an overall motion model independent of body parts, and compares the detected result with parameters in the model by analyzing the overall motion model, to further judge the human body movement state. Compared with the prior art, the method of the invention has a clever design concept, a scientific detection principle, a simple detection method, a high detection accuracy, and a great market prospect.

Description

Technical field: [0001] The invention belongs to the technical field of video surveillance, and relates to a method for detecting body conflict behaviors, in particular to a method for detecting body conflict behaviors based on low-dimensional spatiotemporal feature extraction and theme modeling. Background technique: [0002] In recent years, with the increase of various security emergencies and the improvement of public security awareness, along with the penetration of artificial intelligence concepts and the continuous maturity of artificial intelligence technology, intelligent monitoring has attracted more and more attention. The traditional monitoring system mainly realizes the safety management of public places through manual monitoring, which lacks real-time and initiative. In many cases, video surveillance only plays the role of video backup due to unmanned management, but does not fulfill the responsibility of supervision. In addition, with the popularization and w...

AI Technical Summary

Problems solved by technology

Existing inventions for abnormal behavior detection have the inherent characteristics of failure to truly understand abnormal behavior, so the existing abnormal behavior detection model cannot fully reflect the essence of abnormal behavior, resulting in the detection accuracy obtained based on the existing abnormal behavior detection model. The ideal effect was not achieved. Therefore, a detection method for body conflict behavior based on low-dimensional spatiotemporal feature extraction and topic modeling was designed. The calculation method is accurate and the detection result is accurate.

Images