Biomechanical systems such as horse locomotion are investigated by using inertial sensor systems composed of accelerometers, gyroscopes and magnetometers, or by optical motion tracking systems. The major difficulty in the inertial sensor systems is the integration process. Furthermore, the signals possess noise that has to be filtered. On the other hand, the motion tracking systems are expensive and mostly adapted to indoor laboratory conditions. Hence, in some studies, horses are trained to walk on treadmills with various speeds. This training process is time consuming and may lead to unnatural gaits. In this study, an inexpensive, portable vision system for tracking the motion in a large calibration volume is designed; and an algorithm for obtaining kinematics of a rigid body is developed and implemented in MATLAB. Using the overall system, it is possible to determine the position, velocity and acceleration of any point in the calibration volume which may contain multiple rigid bodies. The point of interest may correspond to points which do not allow an accelerometer to be mounted, or points which are invisible to the cameras. This is the major advantage of our method. A singularity analysis of the algorithm, which yields useful information on the positioning of the markers that are to be tracked by the vision system, is also performed. The method is used to study the whole-body vibrations (WBV) imposed on a horse rider. To the authors' knowledge, there exists no studies on this topic in the literature. Regarding the imposed WBV, the differences between the gaits of the same horse and the differences between the gaits of different horses (executing the same gait) are investigated. The results are experimentally shown to be consistent with the comfort assessment of four experienced riders.