Conventional histologic methods provide valuable information regarding the physical nature of damage in fatigue-loaded tendons, limited to thin, two-dimensional sections. We introduce an imaging method that characterizes tendon microstructure three-dimensionally and develop quantitative, spatial measures of damage formation within tendons. Rat patellar tendons were fatigue loaded in vivo to low, moderate, and high damage levels. Tendon microstructure was characterized using multiphoton microscopy by capturing second harmonic generation signals. Image stacks were analyzed using Fourier transform-derived computations to assess frequency-based properties of damage. Results showed 3D microstructure with progressively increased density and variety of damage patterns, characterized by kinked deformations at low, fiber dissociation at moderate, and fiber thinning and out-of-plane discontinuities at high damage levels. Image analysis generated radial distributions of power spectral gradients, establishing a "fingerprint" of tendon damage. Additionally, matrix damage was mapped using local, discretized orientation vectors. The frequency distribution of vector angles, a measure of damage content, differed from one damage level to the next. This study established an objective 3D imaging and analysis method for tendon microstructure, which characterizes directionality and anisotropy of the tendon microstructure and quantitative measures of damage that will advance investigations of the microstructural basis of degradation that precedes overuse injuries.