The ever-expanding urbanization and the advent of smart cities need better crowd management and security surveillance systems. Advanced systems are required to improve and automate the crowd management system. The aim of the closed circuit television and visual monitoring systems using multiple cameras faces many challenges like illumination variance, occlusion and small spatial-temporal resolution, person in sleep, shadows, dynamic backgrounds, and noises. Therefore, the crowd monitoring, prevention of stampedes and crowd-related emergencies in the smart cities are major challenging problems. In this paper, we propose an intelligent decision computing based paradigm for crowd monitoring in the smart city. In the intelligent computing based framework, the optimization algorithm is applied to compute the feature of crowd motion and measure the correlation between agents based motion model and the crowd data using extended Kalman filtering approach and KL-divergence technique. The proposed framework measures the correlation measure based on extracted novel distinctive feature, and holistic feature of crowd data represent and to classify the crowd motion of individual. Our experimental results demonstrate that the proposed approach yields 96.20% average precision in classifying real-world highly dense crowd scenes.

Introduction

Crowd management in the smart city is getting more proliferation due to its widespread of application and usage recently. Pedestrian crowd are the essential part of smart cities. To provide better solutions and services in the smart cities, crowd monitoring, planning, and crowd management are necessary [1]. Therefore, various mathematical simulations, theoretically based models, and efficient simulation tools as well as various intelligent support systems using computer vision approaches, pattern recognition and image processing for the crowd management plays a vital role in monitoring and tracking of crowd motion [2]. In the smart cities, different surveillance systems are deployed to monitor the various activities and control and monitor the traffic in the several crowds at several places, such as shopping mall, traffic signals, roads, railways, and airport platforms, etc. The control and monitoring of crowd are important task and major changeling problems in the smart cities recently throughout the world [3]. Nowadays, crowd management and advance visual surveillance system of moving crowds in the smart cities have achieved much attention by scientific interests, research communities, and multidisciplinary researchers to improve crowd planning, management and monitoring of crowd and crowd safety in public facilities at the different places in the smart cities [3] [4]. Deep perspicacity into crowd dynamics and its management have contributed a better platform to several researchers and crowd control researcher groups for the understanding of individual behaviour, tracking of crowd motions in the smart cities. Based on the current approaches crowd consists of discrete individuals able to react with their surroundings [5]. In current years, numerous multidisciplinary researchers, scientists, and research committees have begun to understand and study the management of crowd by detection and tracking of individual behaviours in the crowd [7] [8]. The crowd management system monitors the crowd of smart cities by tracking and identifies the person in the crowds which is already started to prevent the catastrophic events. Therefore, it needs to design and develop automatic crowd management systems and better prediction of the crowded traffic flow in the smart cities across the world. The study of crowd motion is an NP-hard research problem in the computer vision because the individual trajectories of crowd movement are tough to predict the exact individual motions and different complex motions of individual are involved in it. The problem of crowd analysis becomes complicated due to changing nature of crowd density and background scenarios. In computer vision, some of the previous works have been done for the management of crowds as event recognition and anomaly detection [1] [2] [4] [5] [8] [9] in the smart city. However, these monitoring approaches failed to perform the correct prognostication and matching of events with corresponding stored crowd datasets. Other related works based on learning models are applied for evaluating of interactions amongst a small number of persons to analysis the action [7] [10].