Public security has increasingly become a crucial problem world widely ever since “911”. To help relieve this social problem from the technological aspect, engineers have installed a number of cameras in the public areas such as parks, casinos, railway stations etc., expecting to extend the eyes of the security guards. Nevertheless, as the number of regions of interest increases, it results in a huge array of monitoring displays, and thus the human security guards will feel overburdened, distracted and make crucial mistakes in the key surveillance tasks. It is evident that not only the extension of eyes but also the extension of brains is required for public surveillance, so that the system will be capable of some intelligent judgment and detection in the numerous image input and providing strategic feedback. Based on this motivation, crowd modeling technology has been under development to analyze the image input which is constantly crowded with humans, as well as to provide alert signals when abnormal motions emerge.
Instead of solving the problem using heuristic rules, another approach of learning from demonstration examples is used. The complexities of many human actions make them very difficult to
be handled analytically or heuristically.
We will do the research on four major areas:
(1) Static crowd modeling in term of densitydistribution;
(2) Mathematical description of dynamiccrowd modeling using velocity field theory;
(3) Understanding of human crowd motions(e.g. gathering, dispersing, etc.) and behaviors (e.g. cheering, panic, etc.);
(4) Detection of individual’s abnormal behaviors in the crowd.
This work has potential applications in different areas of our everyday life. First of all, it can be applied to intelligent surveillance systems in security agencies. Secondly, this research will enhance the intelligent level of transportation systems in areas such as evacuation, crowd guidance, etc. The understanding of crowd distributions and motions will make crowd control easier and more skillful. The schedule of public transportation vehicles such as trains and buses
can be optimized to provide larger load. Thirdly, for the tourism agencies, crowd modeling and monitoring technology can be employed to optimize the controlling and guiding of tourists flow for safety, comfort, and protection of resources.
|
