This paper presents a novel method for interpreting data to improve the indirect structural health monitoring (SHM) of bridges. The research presented in the study is part of an ongoing study aimed at developing a novel SHM paradigm for the health assessment of bridges. In this paradigm, we envision the use of an instrumented vehicle that assesses a bridge's dynamic characteristics while traveling across the bridge. These characteristics are then correlated to the health of the structure by means of advanced signal processing and pattern recognition approaches. In this paper, we present and compare two classification algorithms that locate the presence of damages at well-defined locations on the structure: sparse representation and the Fourier discriminant methods, and find that the sparse representation method provides superior classification accuracy.