The topic of this thesis is lifelogging, the automatic, passive recording of a person’s daily activities and in particular, on performing a semantic analysis and enrichment of lifelogged data. Our work centers on visual lifelogged data, such as taken from wearable cameras. Such wearable cameras generate an archive of a person’s day taken from a first-person viewpoint but one of the problems with this is the sheer volume of information that can be generated. In order to make this potentially very large volume of information more manageable, our analysis of this data is based on segmenting each day’s lifelog data into discrete and non-overlapping events corresponding to activities in the wearer’s day. To manage lifelog data at an event level, we define a set of concepts using an ontology which is appropriate to the wearer, applying automatic detection of concepts to these events and then semantically enriching each of the detected lifelog events making them an index into the events. Once this enrichment is complete we can use the lifelog to support semantic search for everyday media management, as a memory aid, or as part of medical analysis on the activities of daily living (ADL), and so on. In the thesis, we address the problem of how to select the concepts to be used for indexing events and we propose a semantic, density- based algorithm to cope with concept selection issues for lifelogging. We then apply activity detection to classify everyday activities by employing the selected concepts as high-level semantic features. Finally, the activity is modeled by multi-context representations and enriched by Semantic Web technologies. The thesis includes an experimental evaluation using real data from users and shows the performance of our algorithms in capturing the semantics of everyday concepts and their efficacy in activity recognition and semantic enrichment.