Displacement measurement has always played a critical role in improvements in technology. As there are various needs of measurements, a variety of type of sensors has been developed. For displacement measurements in very narrow spaces, such as observing the deformation of vehicle components during test driving, the space and surfaces available for measurement is limited. Among many types of displacement sensors, vision-based displacement sensors are suitable for such circumstances. Vision-based displacement sensors are mainly utilized in measuring slow displacements from a far distance, such as structure health monitoring. This is due to the high computational power required for fast data acquisition in vision-based sensors. In order to extend the applications of vision-based displacement sensors, increase in the data acquisition rate is necessary. This paper will propose an algorithm using marker detection and image cropping to minimize the amount of calculations required to extract the displacement information from images, enabling real time vision-based micro-dynamic displacement measurement.