Abstract
As reported in the latest plans of Hong Kong government, there is a pressing need for Hong Kong to be an intelligent city. Our group is going to take on this task and make Hong Kong a really intelligent city. Therefore, the main purpose of this project is to automatically detect or even predict any emergency incidents that happen on the roads (such as car accidents, broken water pipes, and other road blocking incidents caused by football matches, protests and so forth), and then alert drivers and provide alternative routing solutions to them.
Our project is currently divided into two parts: data collection and data analysis. For data collection, we will consider write an android mobile application to collection necessary geographic information from users. Specifically, the data collected should contain latitude, longitude, altitude and the timestamp. Given these data, we are going to convert it into three-dimensional coordinates for the sake of accuracy and convenience. The three-dimensional coordinate system will be built up accordingly after some experiments. For data analysis, after our literature research(see references), we found that there are mainly four methods to detect traffic accidents:They are distance-based, deviation-based, distribution-based and density-based methods. At this stage, we consider to focus on the first two methods.