Vibration measurement is of great significance for the modal recognition, structural safety monitoring and operation reliability evaluation of large-scale wind turbines blades. Traditional contact sensors are limited by full-field measurement and insufficient real-time modal parameter identification ability. In this study, a non-contact vibration measurement method integrating Digital Image Correlation (DIC) with Covariance-Driven Stochastic Subspace Identification (COV-SSI) is proposed, and a cantilever beam is adopted for experimental verification. A transient excitation is applied using a force hammer, and the vibration image sequences are synchronously collected by a stationary camera. The full-field dynamic displacement response is calculated based on the DIC algorithm. Further-more, COV-SSI is utilized to identify modal parameters from the time-domain response. The results are compared with data collected by accelerometers. The comparison shows that the relative error of the first-order natural frequency obtained by the proposed method is less than 0.3%, along with a relative error of less than 1.1% for the corresponding damping ratio. The agreement verifies the effectiveness and practicability of the non-contact method in structural vibration monitoring.