Detecting the initiation of mechanical injury of soft tissue, and not only its ultimate failure, is critical to enable a sensitive and specific characterization of tissue tolerance, to develop quantitative relationships between macro- and micro-structural tissue responses, and to appropriately interpret physiological responses to loading. We have developed a novel methodological approach to detect the onset and spatial location of structural damage in collagenous soft tissue before its visible rupture, via the identification of atypical regional collagen fiber kinematics that were produced during loading. Our methods utilize high- speed quantitative polarized light imaging and identify the onset of tissue damage in ligament regions where the mean collagen fiber rotation significantly deviates from its behavior during previous noninjurious loading. This technique was validated by its ability to correctly predict the location of visible rupture (p<0.001). The proposed fiber rotation-based metric identifies potential facet capsular ligament damage beginning well-before rupture, at 51±12% of the displacement required to produce tissue failure. While traditional macro-scale strain metrics fail to identify the location of microstructural damage, detection determined by altered fiber rotation is significantly correlated (R=0.757;​ p=0.049) with the occurrence of tissue yield, supporting the capabilities of this method. Damaged regions also exhibit higher variance in fiber direction compared to undamaged regions (p=0.041). These methods enable the characterization of a previously undefined class of mechanical injuries to the facet capsule, and have potential utility to define more refined injury tolerances and to provide region-specific and fiber kinematic data for finite element and tissue-level models.