ABNORMAL CROWD BEHAVIOUR DETECTION BASED ON THE HABITAT LOCATION
V. Deepiga
, Mrs.S.Revathy
Video, Crowd density, MATLAB
Crowd density estimation and abnormal crowd behaviour detection have become increasingly important subjects of research in computer vision. Some of the issues in crowd analysis include perspective distortion, occlusions, illumination changes, etc. Crowd density estimation and crowd behaviour detection plays a vital role in security systems, congestion control and public space design. We propose a novel framework for crowd density estimation and abnormal crowd behaviour detection. In crowd density estimation, we propose a novel feature extraction method by combining the texture and pixel features of crowd. These features are then given to a neural network to estimate the level of crowd density and perspective distortion has also been handled. In abnormal crowd behaviour detection, we propose a novel edge based feature points extraction method that eliminates the need to remove the static feature points in the subsequent frames. KLT tracker has been used to track the points and to determine the principal direction of motion of each feature point a direction count algorithm is proposed. The motion attributes such as principal directions, speed are used to model the crowd behaviour based on threshold values. Experimental results show that our proposed method identifies the crowd density level with accuracy 95% and crowd behaviour with accuracy 97%.
"ABNORMAL CROWD BEHAVIOUR DETECTION BASED ON THE HABITAT LOCATION", IJSDR - International Journal of Scientific Development and Research (www.IJSDR.org), ISSN:2455-2631, Vol.2, Issue 3, page no.210 - 215, March-2017, Available :https://ijsdr.org/papers/IJSDR1703032.pdf
Volume 2
Issue 3,
March-2017
Pages : 210 - 215
Paper Reg. ID: IJSDR_170119
Published Paper Id: IJSDR1703032
Downloads: 000347175
Research Area: Engineering
Country: -, -, India
ISSN: 2455-2631 | IMPACT FACTOR: 9.15 Calculated By Google Scholar | ESTD YEAR: 2016
An International Scholarly Open Access Journal, Peer-Reviewed, Refereed Journal Impact Factor 9.15 Calculate by Google Scholar and Semantic Scholar | AI-Powered Research Tool, Multidisciplinary, Monthly, Multilanguage Journal Indexing in All Major Database & Metadata, Citation Generator
Publisher: IJSDR(IJ Publication) Janvi Wave
