The goal of the research described here is to present an approach for automating the detection and the extraction of meaning from text using a range of linguistic and ontological techniques, concepts such as the lexico-semantic functions proposed in Meaning-Text Theory by Mel?cuk and the concept of the context. This is motivated, by the fact that, on one hand, these functions enable a better modeling and formalization of meaning. On the other hand, the concept of the context is not well supported in many important and essential domains such as in Natural Language Processing (NLP), semantic data analysis and knowledge management. Our work consists here to integrate these elements in an automatic and context-based approach we called PROSEM (SEMantic PROjection), based on an extended model of the context, that improves the detection of meaning in a text. This approach can be used in many areas related to the meaning and the context such as semantic data indexing, linguistic or semantic search, contextual information extraction, multilingualism, etc. In this paper, we present the main elements of the approach and three typical applications processes enhanced by PROSEM. An estimate of the performances of the approach will conclude the article.