With the burgeoning of social network sites, social image tagging, where images are annotated with tags, is gaining traction. But, in the domain of social media, this task becomes complex as images are associated with plenitude of object information. In this paper, we have put forth a Hybrid Classifier-based Ontology driven image Tag Recommendation framework for social image tagging which is based on a knowledge-centric, metadata generating, semantically-inclined and ontology-driven strategy. The COIL-100 dataset is used for the implementation. The proposed HCOntoTR approach is evaluated using sundry performance measures by comparison with baseline models like UIT, HRSDL and RSSVM. The proposed HCOntoTR model gives finer results set against the baseline models, with a precision of 94.49 % which surpasses the others, owing to the amalgamation of a profusion of knowledge from heterogeneous knowledge sources namely the Wikidata and Google's Knowledge API.