Coronal Holes (CH) are large-scale, low-density regions in the solar atmosphere which may expel high-speed solar wind streams that incite hazardous, geomagnetic storms. Segmentation of CH boundaries may aid in validating predictive coronal and solar wind models but has proved difficult due to similar appearances as filaments and the tendency for dense coronal plasmas to block underlying CHs in Extreme Ultraviolet (EUV) imagery. We propose an automated detection algorithm of CHs which revisits ground-observed, chromospheric He I 10830 Å line imagery and underlying photospheric magnetograms as drivers to circumvent these issues and provide a complementary method to the space-observed, coronal EUV emission-driven methods that have been widely adopted in the community. Classical computer vision techniques are applied to imbue the routine with design variables based in radiative and magnetic properties of CHs, as well as produce an ensemble of boundaries with quantified intra-algorithm uncertainty. This method is science-enabling towards future studies of coronal hole formation and variability from a mid-atmospheric perspective.