Motion capture is the process of recording the three-dimensional (3D) motion or positioning of a subject and/or object in a global (or laboratory) reference frame. There are a number of different technologies that can be used to either directly measure or to estimate the 3D motion. Motion capture traditionally refers to technologies that directly measure 3D movements from markers or sensors. I prefer the term motion tracking as there are other technologies that estimate 3D movement from inertial measurement units (IMU) without directly capturing the motion.
The main technologies used for motion tracking are:
A basic understanding of how motion tracking works is necessary to compare the different technologies. Most people associate traditional motion capture with a subject having reflective markers on their body and the resultant 3D positions of the markers in the global reference frame are calculated from optical cameras. That is where the term motion capture comes from, as the 3D motions of individual markers are directly measured with the optical camera systems. By using at least 3 non-collinear markers on each body segment, the 3D position of each joint segment can be calculated. Similar results can be obtained with a magnetic system, which directly measures joint segment 3D motions. So both optical camera systems and magnetic systems can directly measure the 3D motion of a subject or an object. From the direct measurements, the motion data can be differentiated to obtain velocity results, and then differentiated again to obtain acceleration results. This is a relatively trivial mathematical problem to perform with modern day computer software algorithms. However, any errors in 3D positions will be amplified with each differentiation step.Another way to provide motion tracking is to use IMU technology. Rather than directly measuring 3D positions in a global reference frame, these devices directly measure local accelerations with accelerometers and angular velocities with rate gyros. As the figure above shows, velocity can be obtained by integration of the acceleration components, and 3D positions can be obtained by integration of the velocity components. Integration involves a more difficult mathematical algorithm as we need to know some global reference points in order to perform the integral calculus needed to convert the local measured acceleration and velocity data into the global (laboratory) frame. By using magnetometers or GPS technologies, advanced algorithms can be used to estimate the global 3D positions of the body segments from direct measurement of local acceleration and angular velocity data. So while IMUs do not directly capture motion data, they can accurately track the motion data.
Understanding these motion tracking basics are important in order to be able to compare results from different system technologies, or more importantly to determine which systems are better for different applications. While they all provide the same relative information, there are pros and cons for each technology and the accuracy of each of the kinematic variables (positions, velocities, and accelerations) will vary dependent upon what is directly measured and how many calculations or algorithms are required to calculate the other kinematic variables.
Having extensive experience with each type of motion tracking technology, I am very familiar with the pros and cons of each technology. I have provided a lot of that information in the tabs for each type of technology (optical, magnetic, and IMU).