Most of the displacement and vibration testing methods based on vision require artificial marker on structure surface, which makes vision-based testing and monitoring for real-life more complicated. Aiming at the disadvantages of the existing machine vision technology in vibration modal testing for thin-walled components, a vibration modal testing method of thin-walled components based on curvature scale space (CSS) corner detecting and matching is proposed. The virtual corner is set to replace traditional physical targets, and the key point of vibration image sequence of thin-walled components is oriented by the improved CSS corner detection algorithm, which is matched by using. Thus, the vibration information is obtained and modal parameters are identified. Eventually, the vibration modal testing experiments for thin-walled beam and thin-walled cylinder is carried out. By analyzing measurement results and verifying with finite element simulation and accelerometer, the results show that the proposed method can effectively and accurately obtain modal parameters, and the measurement error is less than 5%.