Blind digital image forensics, aims at identifying the authenticity and origin of the digital images without any embedding information beforehand, has become one of the research hot spots recently. Resampling, as a kind of tampering manipulation, will change the statistical features of the images. In this paper, the specific statistical changes brought into the linear dependencies among image pixels are analyzed by singular value decomposition. A blind and efficient method capable of finding traces of resampling and interpolation based on singular value decomposition is proposed. The performance of resampling detection shown in experimental results has indicated the validity of the algorithm.