CROWD SIZE ESTIMATION USING MATLAB

Abstract : A crowd is something beyond a simple sum of individuals. It has collective characteristics which could be described in general terms such as ‘angry crowd’, ‘peaceful crowd’. A crowd can assume different and complex behaviors as those expected by their individuals. Understanding crowd behavior helps in designing pedestrian facilities, for major layout modifications to existing areas and for the daily management of sites subject to crowd traffic. Conventional manual measurement techniques are not suitable for comprehensive data collection of patterns of site occupation and movement. Real time monitoring is tedious and tiring but safety-critical.

 EXISTING SYSTEM :

 The Fitness of each chromosome (which represents an image), is defined as a subjective fitness score between 0-10. It is here where human intervention occurs, since it is a human interpreter that assigns the fitness of each individual chromosome. The interpreter may set the fitness based on the brightness and enhancement of certain areas of an image. The result of this, is that the final output image will be enhanced in a manner that the user desires. This operation would be too human intensive if every chromosome was to be considered, thus a method is developed whereby the human interpreter only looks at a subset ( ) of the chromosomes from the total population size(N).

 DISADVANTAGE :

 The up and down oscillatory head movement of individuals walking in a freely flowing crowd stop when the crowd is too dense for free movement. We compute the 2-dimensional Discrete Fourier Transform (DFT) for each image in a time sequence followed by a measurement of temporal changes in the resulting magnitude and or phase spectra. However this approach has two main disadvantages: a) DFT for a single image is related to local changes of intensity and not to temporal interframe properties. b) It involves a high computational and memory cost

 PROPOSED SYSTEM :

 The other methods suffer from near far effects where people near to the camera will occupy more area than people of the same size far from the camera. This is exacerbated if standard closed-circuit security camera installation is used for data-capture, because such cameras are commonly mounted at low angles either to obtain a longer field view for human monitoring or because they are installed in locations with low ceilings. For instance tunnels or platforms in subway areas.

 ADVANTAGE :

 1. Grey-level segmentation or Threshold method. 2. Edge- Detection Techniques. 3. Digital morphology. 4. Texture. 5. Thinning and skeletonization Algorithms.