Pedestrian Detection Technology
The number of urban surveillance cameras is on the rise. A majority of such cameras are installed to prevent intruders, to detect suspicious persons and items, and to protect the peaceful life of the urban population. Since the volume of video data produced by such cameras is too great to monitor manually, automatic monitoring systems are required. Implementing such systems requires the ability to detect pedestrians automatically and effectively. Recently, methods that are robust to partial occlusion have also been proposed.
Vehicle Detection Technology
Information technology-based Intelligent Transport Systems (ITS) greatly contribute to the convenience of transport. Vehicle detection is an essential component of ITS. Recently, systems that improve transport safety or measure traffic volumes have been implemented. The focus of current research is to improve the effectiveness of such systems. More specifically, we are researching edge and color feature-based image processing methods for detecting vehicles of various shapes and sizes. This research will enable the automatic detection of vehicles from camera videos.
Lane Detection Technology
The topic of indoor and outdoor autonomous systems has recently enjoyed much attention from various researchers, and some fundamental functions have already been implemented. Such systems require various detection methods, such as detecting the vanishing point in order to determine driving direction; and estimating the potential driving area in order to detect the traffic lane. It is possible to implement effective vanishing point and lane detection based on physical models.
Soccer Player Tracking
Sports like soccer and basketball are viewed by many people around the world. Highlight videos, which summarize an entire match to a short video sequence, are very common. Therefore, the topic of automatically detecting important game events such as goals and corner kicks has attracted much attention. This requires obtaining the players’ locations. In our research, we automatically track players and obtain their locations by focusing on the color of their uniforms and their movement patterns.