抄録
For the visualization of the displacement vector around the fatigue crack tip during the fatigue crack propagation test, PIV based motion analysis is applied to the motion picture which is acquired by the high speed/magnification camera. PIV is the matching point search technique between the time series of the imagery, originally it is developed for the fluid dynamical analysis. In the study, the subarea based correlation matching are used for the displacement vector estimation using the time continuous image frames. The multiple subareas are selected within the first frame, and the corresponding point of each subarea is searched within the designated search region in the second frame. The vector from the origin of the subarea in the first frame to the corresponding point in the second frame (the corresponding vector) contains not only the displacement vector but the movement vector of the test piece, to extrace the displacement vector from the corresponding vector, the average of the corresponding vector is subtracted from each corresponding vector. To make the subpixel size vector estimation, the correlation coefficient of the subarea matching within the search region are approximated by the quadratic formula and find the maximum point as the matching point. The quadratic approximation is used for the elimination of the error matching results as the non-maximum point or the maximum point out of the search region. For the definition of the subarea size the autocorrelation image of the deginated subarea size are used. In the case of too small subarea size, the isolated point cannot be found the autocorrelation image. One of cause of the matching error is the existence of the similar area nearby the designated subarea. The isolated point of the autocorrelation image shows the non-existence of the similar area. By using this scheme, the suitable subarea size is defined. Through the many fatigue test, the results show us that this displacement vector estimation method works well in the cases of that non-defocusing frames are acquired and there are some spatial feature such as the scratchs. The estimated displacement vectors are visualized as the arrow image and the motion picture for the assistance of the understanding of the spatial displacement distribution.
