Raman systems based on handheld fiber-optic probes offer advantages in terms of smaller sizes and easier access to the measurement sites, which are favorable for biomedical and clinical applications in the complex environment. However, there are several common drawbacks of applying probes for many applications: (1) The fixed working distance requires the user to maintain a certain working distance to acquire higher Raman signals; (2) The single-point-measurement ability restricts realizing a mapping or scanning procedure; (3) Lack of real-time data processing and a straightforward co-registering method to link the Raman information with the respective measurement position. The thesis proposed and experimentally demonstrated various approaches to overcome these drawbacks. A handheld fiber-optic Raman probe with an autofocus unit was presented to overcome the problem arising from using fixed-focus lenses, by using a liquid lens as the objective lens, which allows dynamical adjustment of the focal length of the probe. An implementation of a computer vision-based positional tracking to co-register the regular Raman spectroscopic measurements with the spatial location enables fast recording of a Raman image from a large tissue sample by combining positional tracking of the laser spot through brightfield images. The visualization of the Raman image has been extended to augmented and mixed reality and combined with a 3D reconstruction method and projector-based visualization to offer an intuitive and easily understandable way of presenting the Raman image. All these advances are substantial and highly beneficial to further drive the clinical translation of Raman spectroscopy as potential image-guided instrumentation.
copy citation link