This paper discusses the populating of an ontology dedicated to the phytosanitary surveillance of cotton plants. This ontology was built with the aim of annotating the data from phytosanitary surveillance of cotton in Côte d’Ivoire and facilitating collaboration between actors with different skills. Its implementation led to the creation of a semantic Wiki. Populating an ontology consists in associating concrete instances with the concepts and relations defined in ontology. In the life cycle of ontologies, it is recommended that they be constantly updated and maintained. This task can be difficult, costly and tedious when it is dedicated solely to humans and carried out manually. In this paper, we present a semi-automatic approach to ontology populating based on web scraping. This approach combines natural language processing (NLP) and Hearst pattern extraction techniques to automatically detect relevant instances in documents collected from the web. Implementation of this approach resulted in a satisfactory accuracy rate of 79%.