Prediction of Crowd in Three Levels: a Foot Step to Anomalies Prevention
Deepak P1, S. Krishnakumar2
1Deepak P, STAS, Edappally, Kochi, Kerala, India.
2S. Krishnakumar, STAS, Edappally, Kochi, Kerala, India.
Manuscript received on September 16, 2019. | Revised Manuscript received on 24 September, 2019. | Manuscript published on October 10, 2019. | PP: 846-851 | Volume-8 Issue-12, October 2019. | Retrieval Number: L32341081219/2019©BEIESP | DOI: 10.35940/ijitee.L3234.1081219
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Crowd is a huge number of people meet jointly in a disorganized or unmanageable way. Crowd in any atmosphere may leads to suspicious events. Nobody can predict the crowd and some anomalies might happen in the presence of crowd. So prevention of crowd well in advance is the only remedy to tackle the situation. Advance crowd detection is an important subfield in video surveillance. Prior detection or prevention of crowd has so much of importance while we are considering the present scenarios all over the world. So now-days, an automatic crowd prevention technique is needed for all the countries to protect their land, provide safety for their citizens and law enforcement. Crowd prevention system using manual operators are weak due to many physiological and non-physiological factors but it will provide better performance than automatic system in case of decision making. Many models have been developed so far to detect the crowd automatically. Our system aims to predict the crowd well in advance in three levels and so the automatic system or the operator will get enough time to respond or take a decision. To detect the formation of crowd well in advance, all the human objects in a frame was identified by Gaussian mixture model and object classification, shadow was eliminated and crowd was predicted using the object rectangle model and center vertical line model. The pixel distance between the each rectangles and center line is used to predict the formation of crowd. This paper also gives some suggestions to crowd modelling.
Keywords: Video Surveillance System, Anomalies Detection Crowd Detection, Advance Crowd Detection
Scope of the Article: Regression and Prediction