With continuous advancement in optical, electronics, and computer technology, commercial digital cameras now are equipped with a high pixel resolution at a reasonable cost. Image sequences recorded by these cameras contain both spatial and temporal information of the target object; hence they can be used to extract the object’s dynamic responses and characteristics. This paper presents a nontarget stereo vision technique to measure response of a line-like structure simultaneously in both spatial and temporal domain. The technique uses two digital cameras to acquire image sequences of a line-like structure. It adopts a simple nondimensional length matching approach and the epipolar geometry to establish point correspondences within an image sequence and between two image sequences, respectively. After reconstructing a spatio-temporal displacement response from the two image sequences, wavelet transform is then used to extract the modal characteristics of the structure. The technique is illustrated using two free vibration tests: a steel cantilever beam and a bridge stay cable. Results show that the technique can measure the spatiotemporal responses, natural frequencies, and mode shapes of the two structures quite accurately. This image-based technique is a low-cost and simple-to-use approach that not only can complement current response measurement sensors but also offer special advantages which are hard to obtain using these traditional sensors.