Edge detection is the most fundamental step in terms of the high-level image processing. In order to find an edge detection method that has efficient computation time, sensing to noise as minimum level and attaining the skeleton of image, so that various edge detection methods have been developed in edge detection area. In this study, in order to obtain the angle information of each edge point with a different point of view, the concept of Principal Component Analysis is benefited. Finally, straight and thin edges are derived by linking the edge points to each other based upon the obtained edge information.