Behavior recognition method based on three descriptors of hydrodynamics

Abstract

The invention provides a behavior recognition method based on three descriptors of hydrodynamics. The behavior recognition method based on three descriptors of hydrodynamics comprises the steps: calculating three descriptors of a picture, based on hydrodynamics, and respectively obtaining an advection descriptor, a diffusion descriptor and a pressure descriptor of the picture; integrating thermodynamic diagrams of different descriptors of the same frame into the input data type of the adopted neural network, inputting into the trained neural network, performing fine tuning on the neural network to obtain a model of different crowd behaviors; and by classifying the crowd behaviors through the obtained model, obtaining the detected crowd behavior type, and realizing behavior recognition. Thebehavior recognition method based on three descriptors of hydrodynamics has high accuracy for behavior recognition, can reduce a frame in which no event occurs as a frame in which an event occurs, and can also detect more frames in which events occur.

Description

technical field [0001] The invention belongs to the technical field of computer video processing, and in particular relates to a behavior recognition method based on three fluid dynamics descriptors. Background technique [0002] Crowd behavior analysis is an important research topic in the field of computer vision. It mainly analyzes crowd movement behavior in real scenes. Crowd behavior analysis methods can be divided into two types: whole-based and object-based. The whole-based analysis method regards the crowd as a whole , can classify the crowd as normal or abnormal, or identify pre-defined crowd behavior; while the latter regards the crowd as a collection of different individuals, considers the interaction between pedestrians in the scene, and detects abnormal and normal activities in the scene. [0003] Whole-based crowd behavior recognition methods are often used in high-density surveillance scenarios (see: Mehran R, Moore B E, Shah M.A Streakline Representation of F...

AI Technical Summary

Problems solved by technology

[0005] 1) The moving target in the high-density monitoring scene is small, the features detected and extracted from a single individual are extremely small, and the description power for crowd behavior is not enough;

[0006] 2) The features analyzed after being abstracted into particles lose the continuity of the description of crowd behavior;

[0007] 3) Focus on local movement behavior, without considering the overall impact, and lack the description of the entire group

Images