| Summary: | 碩士 === 國立東華大學 === 電機工程學系 === 92 === The purpose of this thesis is to process, analyze and determine the moving direction of micro-scale displacement of objects.
In computer vision, the movement of objects in vision plane causesthe moving of the image pixel and the measurement of the image pixelmoving is generally called optical flow. Moreover, using the optical flow to detect the variations of the continuous images taken by CCD camera is called optical flow approach. Therefore, the optical flow estimation approach is applied to detect the micro-scale displacement of object through the modified calculation of the optical flow field.
In addition, we improve the Least-Squares Fit Method also in this thesis. The strategy for improving the Least-Squares Fit method is that after the calculation of the optical flow the voting probability operation is used to obtain the better estimation results.
|
|---|
