High-speed cameras are reliable alternatives for the direct characterization of optical trap force and particle motion in optical tweezers setups, replacing indirect motion measurements often performed by quadrant detectors. In the present approach, subpixel motion data of the trapped particle is retrieved from a high-speed low-resolution video sequence. Due to the richness structure of motion diversity of microscopic trapped particles, which are subjected to a Brownian motion, we propose to also use the obtained motion information for tackling the inherent lack of resolution by applying superresolution algorithms on the low-resolution image sequence. The obtained results both for trapping calibration beads and for living bacteria show that the proposed approach allows the proper characterization of the optical tweezers by obtaining the real particle motion directly from the image domain, while still providing high resolution imaging.