Semi-automatic Semantic Enrichment of Raw Sensor Data

One of the more recent sources of large volumes of generated data is sensor devices, where dedicated sensing equipment is used to monitor events and happenings in a wide range of domains, including monitoring human biometrics. In recent trials to examine the effects that key moments in movies have on the human body, we fitted fitted with a number of biometric sensor devices and monitored them as they watched a range of different movies in groups. The purpose of these experiments was to examine the correlation between humans’ highlights in movies as observed from biometric sensors, and highlights in the same movies as identified by our automatic movie analysis techniques. However, the problem with this type of experiment is that both the analysis of the video stream and the sensor data readings are not directly usable in their raw form because of the sheer volume of low-level data values generated both from the sensors and from the movie analysis. This work describes the semi-automated enrichment of both video analysis and sensor data and the mechanism used to query the data in both centralised environments, and in a peer-to-peer architecture when the number of sensor devices grows to large numbers. We present and validate a scalable means of semi-automating the semantic enrichment of sensor data, thereby providing a means of large-scale sensor management.