A novel generic framework of event detection in unmanned aerial videos

Abstract

The research on event detection in unmanned aerial videos has received a large attention in the recent years. Plenty of methods have been proposed in this area. However, there is no agreement on which methods fit well for the application in the unmanned aerial videos and which methods fit well for certain event detection. In this paper, we present a framework for event detection in unmanned aerial videos and proposed a series of events to be detected. We mainly divide these events into two categories according to whether the event subject can be well detected and tracked, and all of them are rule-based and model-based. The framework we proposed will adapt to the event detection task. In rule-based event detection, we will give some design methods by way of examples. And in model-based event detection, we will introduce some popular methods in this area and will give out evaluation methods.