Raman spectroscopy using fiberoptic probe combines non-contacted and label-free molecular fingerprinting with high mechanical flexibility for biomedical, clinical, and industrial applications. Inherently, fiber-optic Raman probes provide information from a single point only, and the acquisition of images is not straightforward. For many applications it is highly crucial to determine the molecular distribution and provide imaging information of the sample. Here, we propose an approach for Raman imaging using a handheld fiber-optic probe, which is built around computer vision-based assessment of positional information and simultaneous acquisition of spectroscopic information. By combining this implementation with realtime data processing and analysis, it is not only possible to create fiber-based Raman imaging, but it is also possible to create an augmented chemical reality image of the molecular distribution of the sample surface in real-time. We experimentally demonstrated that using our approach it is possible to determine and to distinguish borders of different bimolecular compounds in a short time. Because the method can be transferred to other optical probes and other spectroscopic techniques, it is expected that the implementation will have a large impact for clinical, biomedical and industrial applications.