The human being in the environments of modern technology has to endure stresses of many and varied kinds of vibrations [1]. Measuring vibration is an important tool in rehabilitation and biomechanical fields of research. We have proposed image processing as a new method to record and determine the frequency response of human body. The subjects were exposed to whole body periodical vibration while standing on a shaking table. Two digital camcorders were used to capture the motion of colored pencil-dot markers on the skin of human body (forehead) and on the edge of the shaking table. After color spotting each frame, the binary image results were processed using new circle factor criteria proposed in this work, for fast finding circles based on second order statistics. The extracted points were calibrated using our own extended version of the direct linear transformation (DLT) method. We subsequently used Borland Delphi 5.0 language to develop useful software for measuring and analyzing human body vibration. As a result, it was clear that the proposed method was lower noise-sensitive in comparison to accelerometer. In order to investigate the validity of the software, the obtained mechanical impedance of the body were compared with other investigations in literature and showed to be compatible. The main advantage of this method is working with a simple user-familiar hardware with no external device attached to the subject and also a user-friendly-software.
On the mechanical impedance of human body surface