Crowd Estimation Using Histogram Model Classification Based on Improved Uniform Local Binary Pattern

As th e number of the world population increases, varieties of challenges are alsoin the rise. Among others, security is a great concern whenever a large number of people gather in some public premises. Therefore, many public areas such as airports, stadiums, and subways are employing crowd monitoring systems to ensure public safety. These systems normally involve the use of closed circuit television (CCTV) where people and their behavior are monitored continuously by security officers. Because the monitoring task is an arduous task, loss of concentration by these humans is inevitable after some period of time. Moreover the observer’s judgment may be influenced by different situations. In addition, it is an uncomfortable task when the officer needs to check several CCTV images concurrently. In such a case, automatic crowd density estimation can be a solution for control managementand monitoring the crowds. Consequently, developing a solution for estimating crowd density is an interesting field for researchers[1]. Since the last two decades, different kinds of methods have been used for automatic crowd density estimation.These techniques could be categorized into either pixel based or texture-based methods.The former method uses background removal and then edge detection and the latter uses a texture descriptor for estimating the crowd density.