An automatic and interactive large-deformation measurement system based on image processing

This paper describes an image-processing-based measurement system by which an automatic large deformation characterization and interactive force control can be realized. The system is provided by a servo-controlled CCD camera and a PC-based frame grabber. A tracking procedure based on the location of a target point is employed for guiding the camera movement. It is intended to monitor the local deformation in a relatively small moving area containing small globules. The globule positions are then used to determine the surface deformation characteristics. In this system, the momentary principal strains being computed online are immediately used to control the applied force. As an illustration, the present system is applied for observing a long-term homogeneous large deformation in SI-rubber specimens being subjected to a constant uniaxial true stress. Four solid, minuscule, relatively undeformable polystyrene globules are used as markers. A true stress control algorithm is developed based on the isotropy assumption in which the transverse area normal to the loading direction becomes determinable from the surface deformation data. It can be stated that the maximum speed of the true stress control is mainly related to the speed of image processing and analysis, which in turn is related to the chosen number of globules and the performance of the PC used. By using four globules, the true stress readjustment could be accomplished for roughly every 4 s. For a large-strain deformation of SI-rubber, it appears that the isotropy assumption is acceptable, since the resulting mean difference between out-of-plane contraction and in-plane contraction is restricted to about±8 percent.